Talk:Book - Manual of Human Embryology 18

From Embryology





Introduction. — The number of papers upon the development of the blood is large, but the majority of them have been written from the clinical standpoint and they often leave much to be wished for the scientific interpretation of the theme. To these clinical writings we owe a confusing nomenclature of the bloodcorpuscles which, unfortunately, has become current in medical works, although it sins against every morphological principle. It unites forms which are morphologically different and separates forms which genetically and morphologically belong together, as is explained more fully in the note, p. 503, and in connection with the discussion of the development of leucocytes. Under these conditions it becomes unavoidable to discard almost entirely the current nomenclature and to replace it by 'a new one. The new nomenclature is in part taken over from others, in part proposed by nryself. It at least corresponds to the morphological demands.

The following exposition is based chiefly on the investigations of four morphologists, — W. His, 0. van der Stricht, J. Jolly, and F. Weidenreich, — to whom we are indebted for the greater part of our present comprehension of the problem of the blood. Of further importance is the just-published (March, 1909) memoir of Maximow (Arch. f. mikr. Anat., vol. lxxiii, p. 444), who studied the development of blood especially in rabbit einbryos. Ruckert and Mollier, 2 in Hertwig's " Handbuck," have given a detailed account of the early development of the angioblast in all classes of vertebrates. The value of this work is very high, and for that reason we regret very much that they have not included the eytomorphosis of the blood-corpuscles within the limits of their account. Although I am unable in many cases to adopt the point of view of the clinicians as my own, yet I have collected from their writings many data.

1. The Angioblast. — Comparative embryology teaches us that the first bloodvessels appear upon the yolk-sac collectively and at one time. They form a unit anlage, which we call briefly the angioblast, according to the suggestion of His. It must, however, be immediately mentioned that several investigators, like Maximow in his latest paper, derive the blood-vessels directly from the mesoderm of the embryo. In fact, we can assert the complete precocious independence of the angioblast from the mesoderm proper only as highly probable. The angioblast lies originally immediately upon the yolk and forms a network that can be recognized just after the first appearance of the anlage. The mesoderm, sensu strictu, forms a continuous layer which lies above the anlagen of the vessels and comes into direct contact with the yolk only in the gaps of the vascular network. According to the majority of observations the angioblast appears to be split off, in all vertebrates, directly from the yolk. It is very difficult to decide whether the *To Professor Mall I am specially indebted, for he has had the kindness to lend me extremely valuable material from his embryologieal collection.

2 Ruckert und Mollier : Die erste Entstehung der Gefasse und des Blutes bei TVirbeltieren, Hertwig's Hdbch. vergi. Entw. "Wirbeltieren, vol. i, p. 1910-1278.


DEVELOPMENT OF THE BLOOD. 499 angioblast is to be interpreted as belonging genetically to tbe middle germ layer or as a derivative of tbe entoderm. Tbe views as to tbese interpretations are very divergent, but tbe fact remains tbat tbe angioblast becomes independent very early and is tbe first tissue of tbe embryo to exbibit an unquestionable differentiation and sbarp limitation. It must be especially empbasized tbat tbe vascular anlagen do not develop in common witb tbe mesoderm, or, if one prefers, witb tbe remaining mesoderm. I incline strongly to the opinion tbat the mesoderm is formed first and that the angioblast, added later, forms itself, not through transposition and transformation of mesodermic cells already present, but from cells which separate from the yolk, or from the layer of yolk cells, and form a reticulate grouping of themselves between tbe middle and lower germ layers.

The angioblast probably maintains its complete independence throughout life. In other words, it is probable that the endothelium of the blood-vessels (and of the lymph-vessels) and the blood-cells at every age are all direct descendants of the primitive angioblast. Unfortunately, our present knowledge does not allow us to express an opinion on this point with absolute confidence. Thus, we find that Maximow (Arch. f. mikr. Anat., vol. lxxiii, p. 511-515) attributes the formation of new vessels and of new mesamceboids, not to tbe angioblast, but to tbe mesoderm proper. The most recent American observations speak against Maximow's view.

The differentiation of the angioblast in amniota may be summarized as follows : The network consists originally of cell cords, which soon become hollow. According to many observations the cavity may be bounded at first on its under side only by yolk. The angioblast cells transform themselves in part into endothelial cells, in part into new blood-cells. The endothelium arises from the peripheral layer of tbe cords; blood elements, on the contrary, from the more centrally placed cells. Only the endothelium forms an uninterrupted network; the blood-cells form scattered clusters, the so-called blood-islands. These consist of cells which are not separated by cell walls either from one another or from the neighboring endothelium. Very often, perhaps always, one finds the lumen of the vessel below (entad) the blood-islands, the cells of which hang down in a cluster from the upper surface of the vessel. In the majority of amniota, the blood-vessels arise in a limited space, which surrounds the embryo and covers only tbe upper surface of the yolk. This space is tbe area vasculosa. In man, however, the area covers the whole yolk from the start. The area vasculosa, studied in fresh specimens, can be recognized in many amniota by the red color of the bloodislands. This color corresponds to the beginning of the development of haemoglobin. Soon the cells of the islands separate from one another and become free. They are the primary blood-cells, or, better, the primitive mesamceboids. Very often the first-formed cells are quite large ; nevertheless, they possess the ability to wander out from the vessels, giving rise in this way to the giant wandering cells which one can observe in very young embryos, as, for example, those of the chick. The large primary cells become gradually smaller by repeated division until they reach the condition which I regard as the rejuvenated stage of the blood-cells, with which the cytomorpbosis proper begins. The mesamoeboids are round cells with relatively large nuclei, which are approximately round. The nucleus is surrounded by a thin layer of protoplasm which, on account of its slight thickness, has often been overlooked. The nucleus has a distinct reticulum, the nodes of which are thickened in part, forming so-called plasmasomes. The protoplasm is finely granular. Cells with these distinct characteristics occur in all vertebrates, but are restricted to early embryonic stages, and have not, up to the present time, been observed in adults. The cells in question multiply in the blood by mitotic division. Their bodies soon become larger, and thus arise colorless cells which continue to divide. Their descendants develop in different ways, in part retaining the embryonic habitus, and in part transforming themselves into red cells — erythrocytes. It must be further noted that in relatively late embryonic life the undifferentiated mesamceboids in part develop into genuine leucocytes.



The primitive mesamoeboids are the ancestors not only of all blood-cells, but also, as Maximow has demonstrated, of other cell forms which occur in the connective tissue of the adult. Recent morphological investigators of the blood consider the conclusion secure that red and white blood-corpuscles have the same origin, or, in other words, that they arise monophyletically. Especially illuminating are the investigations of Maximow 3 and Frau DantschakofT on this question.

The majority of embryologists are of the opinion that the colorless mesamoeboids remain throughout life in order to serve as a permanent source of both red and white blood-cells. Since they can move freely, they can alter their distribution in the body. In mammals the multiplication of the primitive mesamoeboids during the earliest development occurs only in the yolk-sack; later it takes place in the circulating blood; still later in the fetal liver and lymphoid organs; and, finally, in the marrow of bones, which serves as the permanent site of blood formation. Up to the present time no conclusive proof has been brought that the cells in question arise autochthonously in the liver or lymphoid organs or bone-marrow. Therefore, embryologists incline to the opinion that we have to deal merely with the accumulation of immigrant cells. In other words, according to the present view all the cellular blood elements are direct descendants of the primitive mesamoeboids. That this view is secure beyond all doubt cannot, however, be asserted.

The development of human blood still awaits a thorough investigation. The observations at present available are in great part — though not exclusively — more or less incidental to other researches. We find data, first, in descriptions of the development of certain organs, especially the yolk-sack, the liver, and the bone-marrow; secondly, in more extended articles on blood development. The number of such articles is very large, but they are chiefly occupied with the phenomena as observed in various animals. Schridde has studied the blood development in nine young human embryos and has reached conclusions which cannot easily be brought into agreement with other apparently reliable observations. Unfortunately, his research is known to me only through his preliminary notice of 1907 (Verh. deutsch. pathol. Ges. fur 1907, p. 360-365). Therefore a critical discussion of his work is excluded.

We possess at present two comprehensive memoirs, in which the previous literature — so far as it concerns the red blood-corpuscles of vertebrates — is extensively considered and critically discussed. The memoir of Weidenreich B appeared in two parts, of which the first deals with the form and structure of red corpuscles, while the second describes the immature forms and the origin and transformation of the colored corpuscles. Weidenreich strives to give a unified summary of the results already obtained. The memoir by Jolly 6 offers us not only the results of an excellent comprehensive investigation of the cytomorphosis of the blood-cells, but

3 Maximow : Arch, f . mikr. Anat., vol. lxvii, 1906, p. 680-757, and vol. lxxiii, 1909, p. 444-561; Folia Haematol., vol. iv, p. 611-626; Verhandl. Anat. Ges., vol. xxxii, p. 65-72.

4 Wera Dantschakoff : Entwick. d. Blutes b. Vogeln, Anat. Hef te, vol. xxxvii, p. 471.

6 Franz Weidenreich : Die rothen Blutkorperchen, I, Ergeb. anat. Entw. Ges., vol. xiii, 1905, p. 1-94; II, ibid., vol. xiv, 1905, p. 345-450.

9 J. Jolly : Recherches sur la formation des globules rouges des mammiferes, Arch. Anat. microsc, vol. ix, 1907, p. 133-314.

DEVELOPMENT OF THE BLOOD. 501 also valuable discussions of previous investigations. The views defined by Jolly deserve special attention because they have been worked out very conscientiously. Since an exhaustive consideration of the development of the phenomena in animals lies outside the limits of our present undertaking, it seems suitable to recommend the memoirs of Weidenreich and Jolly as excellent sources for the reader who seeks exact literary data.

2. Origin of the Human Angioblast. — Our knowledge of the actual facts is here very defective. We do not yet know, by actual observation, how the earliest vessels arise in man. We know merely that the angioblast appears first on the yolk-sack, and that in almost the earliest stage known to us it already occupies the whole surface of the sack. The angioblast has genuine bloodislands and grows later into the embryo presumably by the formation of sprouts. The precocious development of the human angioblast is, in all probability, closely connected with the precocious independent development of the yolk-sack. A few of the more exact data may be presented. Graf von Spee 7 observed a few islands in the wall of the yolk-sack in an ovum of 9 mm. diameter, with an embryonic shield of 0.37 mm. The yolk-sack had a diameter of 1.84 mm. Its mesoblastic covering formed irregular bunches and projections, which were especially noticeable around the pole of the sack farthest from the embryonic shield. Each of these eminences corresponded to a blood-island situated between the mesoderm and the entoderm. Keibel 8 observed similar relations in an embryo of 6 mm. X 8.5 mm. (including villi). The vascular anlagen do not occur in the immediate neighborhood of the embryo, but begin at a line somewhat removed from it. The yolk-sack of an embryo of 1 mm. (Harvard Embryol. Coll., No. 825) comprises two territories : one of these I regard as the area pellucida, because it possesses a very thin entoderm and occupies the embryonic half of the sack; the other territory, which I regard as the area opaca, has a thicker entoderm and lies opposite the embryonic shield. It is to be noted, moreover, that the vascular formation is restricted to this second territory. This case, which up to the present is unique, renders it probable that in man also the formation of the angioblast begins in a true area opaca and then spreads out toward the embryo, as occurs typically in other amniota.

It is well known that in many amniota the earliest mesamoeboids are relatively large. They probably arise directly by pinching off from the yolk (entoderm), and soon thereafter they are separated from the yolk, or entoderm, by the growth of the vascular endothelium around them, by which they become enclosed in a definitive vascular space. The large primitive mesamoeboids, isolated in the manner described, multiply quite rapidly and become smaller. Meanwhile the circulation has begun, and at least a part of the mesamcoboids leave their place of origin. The history 'Graf von Spee: Arch. f. Anat. u. Entwicklungs Ges., 1896, p. 8. "Franz Keibel: Arch. f. Anat. u. Entwicklungs Ges., 1S90. p. 255.


of the cells has not by any means become clear to us, for it still remains uncertain whether they all pass through the same transf omiations ; but it is certain that many transform themselves into cells of very small size, with nuclei much smaller than those of the cells in the neighboring germ layers. Their protoplasm is minimum in man, so that the cell bodies form merely thin coverings for the nuclei. This process — multiplication of the nuclei and the retarded growth of protoplasm — is a general phenomenon in the earliest development of metozoa, and I have regarded it " as the rejuvenating process with which ontogeny must begin. If we accept this view, we may say that the mesamceboids rejuvenate much more rapidly than the other cells of the germ layers.

While we must admit that our knowledge of the earliest development of the blood in vertebrates is but little satisfactory, because it does not touch many essential points, we must add that the corresponding processes in man are, properly speaking, unknown to us.

The growth of the vessels into the embryo occurs very early. In embryo Klb. 10 (1.8 mm., 5-6 segments) the vessels have already passed into the embryonic body and lie between the visceral meso

Fig. 354. — Three primitive mesamceboids from the yolk-sac of a human embryo of about 1 mm. Harvard Emb. Coll., series 825. X 1500.

derm and the entoderm. The same pathway between the germinal layers is followed, in all vertebrates, by the first vessels as they grow into the embryonic body. Schridde (I.e., p. 362) found in an embryo of 2.5 mm. a similar net of vessels. In human embryos of 8-10 segments the chief primitive vessels are present. They are of merely endothelial tubes.

3. The Primitive Mesamosboids of Man. — The cells in question have not yet been investigated accurately. 11 In the yolk-sack of the embryo mentioned above (H. E. C, No. 825) there occur cells which I regard as primitive mesamceboids, three of which are represented in Fig. 354. They are characterized by the largeness of their nuclei and the small amount of their protoplasm. The nuclei possess about the same dimensions as the nuclei in the neighboring mesoderm and entoderm. The karyoplasma forms a dense superficial layer and a very wide meshed net of fine threads in the interior, with a few thickenings of varying sizes and irreg

Minot: The Problem of Age, Growth, and Death, New* York, 1908. 'Keibel: Normentafeln, Heft viii, p. 20. Compare Schridde; Verhand. dentsch. pathol. Ges., 1907, p. 360.

DEVELOPMENT OF THE BLOOD. 503 ular distribution. Not infrequently, however, there is a single main thickening centrally placed. The nuclei scarcely differ in structure from those of the neighboring tissues. The cell body is finely granulated, irregular in form, without a membrane, and is more deeply colored than the nuclei. I consider it probable that the cells are amoeboid. Maximow 12 gives a detailed description of the primitive rnesamoeboids in the rabbit, to which the reader is referred because the author goes more into detail than is at present possible for the cells in man.

4. Cytomorphosis of the Erythrocytes. — The erythrocvtes arise from the primitive rnesamoeboids, the cell bodies of which become laden with haemoglobin and at the same time acquire a homogeneous appearance. Meanwhile the nuclei also undergo important alterations.

The development of the erythrocytes has been much studied. In the majority of the published papers one feels the lack of a scientific morphological interpretation of the observations, many authors being interested chiefly in clinical applications.

We can confidently distinguish four chief stages in the cytomorphosis of the red corpuscles, for which I propose the following designations: 1. The rnesamoeboids, the primitive or earliest colorless cells, which appear at first in the blood-spaces and arise chiefly, perhaps exclusively, by the breaking up of the blood-islands.

2. The Erythrocytes. — This term includes all red blood-cells which arise, probably exclusively, from rnesamoeboids. They are characterized by their content of haemoglobin and the homogeneous appearance of their protoplasm. We can distinguish three stages in the genesis of the erythrocytes of mammals: A. The ichthyoid blood-cells, the first form of the genuine erythrocyte, which occurs in all vertebrates and constitutes the permanent form in ichthyopsida. In the amniota, on the contrary, they represent a transitory stage of development. The cells in this stage are characterized by their content of haemoglobin, their homogeneous appearance, and their granular nuclei.

B. The sauroid blood-cells, the second form of the genuine erythrocytes, which may be observed as the second stage in the developmental differentiation of the ichthyoids in all amniota. The cells in this stage differ from the iehthyoids by their smaller average diameter, and especially by their smaller, darkly staining (pyknotie) nuclei. The sauroids are atrophying cells. They represent the permanent form in sauropsida, the temporary form in mammals.

C. Blood-plastids. — These are erythrocytes which have lost their nuclei, and occur only in mammals.

Note. — " Mesamceboid " was originally proposed by me to designate the wandering cells which occur in the middle germ layer. The mesamceboid cells, which serve as the parent cells of the red blood-corpuscles, have been often confused with genuine leucocytes, and this has hindered the progress of hematology. The expressions " ichthyoid " and " sauroid " are in themselves not new, but the proposed application of them is new. The term " erythroblasts " has often been used in the sense of our ichthyoid cells, although in these we have to do with red blood-cells already differentiated. Current usage frequently restricts the term to the embryonic forms of the corpuscles in mammals. The mature red corpuscles of amphibia, for "Maximow: Arch. f. mikr. Anat., vol. lxxiii, 1909, p. 464.


example, no one ventures to designate as erythroblasts, although they are homologous with the so-ealled erythroblasts of mammals. " Erythroblast " was introduced by Lowit to designate the colorless cells which, as the preliminary stage of red cells, are appropriately so called. For the regrettable misuse of the word the clinicians are alone responsible. " Normoblast " corresponds to the sauroid cell, but is not always applied with exactly the same meaning. The choice of the term is unfortunate : first, because it seems meaningless from the comparative standpoint, and therefore unavailable; and second, because even from the present clinical point of view it is without significance. The special stage of the " normoblast " is neither more nor less normal than the earlier and later stages. Further, since the stage in question is the permanent one in reptiles, the use of " blast " is unsuitable. " Erythrocyte " is a fitting name for all red blood-corpuscles whether they are nucleated or not, whether their nuclei are pyknotic or not. The effort of the clinicians to restrict this name to the non-nucleated blood-cells of mammals can hardly be justified. At the present day one would hardly expect that red cells with nuclei should not be recognized as erythrocytes but that they should change into erythrocytes by the loss of their nuclei. It does not appear scientific to call a cell Kvrog only after it has become non-nucleated. Similar considerations apply against the use of the expressions " megaloblast " and "microblast." It may be pointed out that in all tissues variations in the size of cells are encountered, and if all these variations are to be specially named the result will be an unlimited confusion in biological nomenclature. I proposed, in 1890, to call the non-nucleated blood-corpuscles of mammals " plastids." At that time I was influenced by the hypothesis proposed by Ranvier, Sehaefer, and others, of the intracellular origin of the red blood-corpuscles. In spite of the fact that the progress of our knowledge has compelled us to give up this hypothesis, we may still term the nonnucleated corpuscles plastids, since the word refers to the fact that they consist of cytoplasm. This renders it possible to ultilize " erythrocyte " as a collective term for any and all red cells, as is done in the present chapter.

The essential characteristic of erythrocytes is haemoglobin, the formation of which may be initiated earlier or later during the development of the single cell. It has long been known that the deposit of haemoglobin may begin in the blood-islands of the area opaca. This phenomenon appears clearly in the sauropsida and has also been recognized in various mammals. In man, on the contrary, if red blood-islands occur at all, they must break up very early; as indeed, according to Maximow, 13 occurs typically in mammals, which in that respect differ from the sauropsida. In the youngest stage yet observed, the free human mesamoeboids do not have any haemoglobin.

We must assume that in man also the primitive mesamoeboids multiply, and that a part of them retain the primitive habitus. While this goes on, one observes the gradual disappearance of the forms with minimal protoplasm. At the end of the first month, and from then on to birth, we find colorless mesamceboid cells of the most varying sizes in the blood-spaces and in the blood-forming organs (Figs. 357 and 359).

M A. Maximow : Arch, f . mikr. Anat., vol. lxxiii, 1909, p. 461.

DEVELOPMENT OF THE BLOOD. 505 Note. — The genetic relations of these cells to one another have still to be accurately determined. In the sauropsida there arise at first very large cells. On the other hand, we learn that the youngest cells multiply and at the same time enlarge. Further, the question arises, Are the large cells in vertebrate embryos all ancestors of the smaller cells or not? We may assume that at least a part of the cells are such ancestors.

Now, while the embryonic blood formation is going on, we may observe, especially in younger embryos, that the erythrocytes differ much in size. In later stages, as in the adult, we find that the developing erythrocytes are much more uniform. From these relations we draw the conclusion that during the developmental period both larger and smaller mesamceboids transform themselves directly into colored corpuscles. But in this connection we must not forget that deductions are less conclusive than direct observations.

The appearance of haemoglobin causes a diffuse coloration of the protoplasm, which at the same time loses its granular appearance and becomes optically homogeneous. Since cells may be observed with varying intensities of coloration, we conclude that there is a gradually increased haemoglobin content of the cell.

Note. — Giglio-Tos maintains that in all vertebrates the haemoglobin arises from special granules. Weidenreich (1. c, p. 406) declares that these granules are artefacts. According to his opinion, we must assume that the haemoglobin appears diffusely in uniform concentration, without being demonstrable in the body of the cell by any special morphological structure. We cannot yet decide whether the haemoglobin is an exclusive or partial product of nuclear activity, as some have supposed, or not.

The cell membrane is probably developed at the same time as the haemoglobin. At least we observe that as soon as the coloration is recognizable the periphery of the protoplasm is bordered by a distinct line. How the membrane is developed is unknown. The nuclei undergo definite alterations during the formation of the haemoglobin, in consequence of which the cells pass to the ichthyoid type. Unfortunately, these alterations have not yet been carefully investigated.

The accompanying pictures represent some of the corpuscles from the blood, — A, of an embryo of 4 mm. (Fig. 355) ; B, of 7.5 mm. (Fig. 356) ; and C, of 9.4 mm. (Fig. 357). In comparison with the earlier stage (Fig. 354) the diminution of the nucleus at once attracts attention.

The more intense nuclear coloration of the older cells is very noticeable. The wide, clear meshes of the nuclear reticulum can no longer be seen. On the other hand, the granules are more numerous, are more deeply colored, and more rounded than before. It is further to be pointed out that there is a striking increase of the cortical layer of the nucleus. These observations may easily

506 be repeated on other embryos. It seems probable that, together with the diminution of the nuclei, increase of the chromatin occurs. This fundamental question, however, cannot be decided on the basis of our present knowledge.




Fig. 355.

Fig. 356.

Preserved with formalin,

Fig. 355. — Two blood-corpuscles of a human embryo of 4 mm. X 1500. colored with alum-cochineal and orange G. Harvard Emb. Coll., No. 714.

Fig. 356. — Three blood-corpuscles of a human embryo of 7.5 mm. X 1500. Zenker's fluid, carmine coloration. Harvard Emb. Coll., No. 256.

In consequence of the changes above described, the cells reach the ichthyoid stage of their development. We have to deal not with the metamorphosis of single cells, but with a genuine cytomorphosis, since the cells continually multiply by division not only during the transformation of the mesamceboids, but also while in the ichthyoid stage. Evidently the cytomorphosis goes on through successive generations of cells.


Fig. 357. — Three blood-corpuscles from a human embryo of 9.4 mm. X 1500. Miiller's fluid, alum-cochineal and safranine. Harvard Emb. Coll., No. 259. Li. nucleus of liver-cell for comparison.

The multiplication of young blood-cells by division was ascertained by Eemak in 1850, and since then has often been observed in many different vertebrates. 14 Mitoses of the ichthyoid bloodcells in the blood-vessels may be observed easily in well-preserved young human embryos. In embryos of 12 mm. the formation of " J. Jolly has published an important paper on the division of bloodcorpuscles in amphibia (Arch. d'Anat. microsc, vol. vi, 1904, p. 455). Jolly gives exhaustive consideration to the literature of the subject.

DEVELOPMENT OF THE BLOOD. 507 the blood in the liver has begun, and after this one finds either no or only exceptional mitotic red corpuscles in the blood-vessels of the body. The blood mitoses of man have not yet been studied in detail.

Bizzozero, 15 after repeatedly studying the multiplication of young erythrocytes, came to the conclusion that after very early embryonic stages the multiplication is accomplished exclusively by the division of cells already containing haemoglobin, and in accordance with this view he denied the continued transformation of colorless cells (Lo wit's erythroblasts) into colored cells. We cannot at present admit that Bizzozero was right.

The sauroid blood-cells arise by the transformation of single ichthyoids. Since, so far as is known at present, they do not multiply by division, they can increase in number only by the metamorphosis of the younger cells. The ichthyoid cells contain hasmoglobin and have a membrane, hence the further visible multiplications concern chiefly the nucleus. There occurs a steady diminution of the volume of the nucleus, and at the same time the framework of chromatin condenses and thickens ; the granules or so-called nucleoli — of which the typical ichthyoid cell has several — become larger and merge with the condensed reticulum so as to become no longer observable (Weidenreich, I.e., p. 407), Fig. 358. The nucleus meanwhile becomes smoothly round, as in other mammals. In this condition it absorbs the usual coloring fluids so intensely that little or nothing can be seen of its structure (Fig. 358). Since the cell does not shrink with the nucleus, the haemoglobin gains the space which the nucleus loses. In brief, the ichthyoid cell changes into the sauroid by pyknosis of the nucleus.

The " normoblasts " of Ehrlich are sauroid cells, but the sauroids vary much in size, a fact which Ehrlich has already pointed out (compare note, p. 504). He directed special attention to the larger and smaller forms, and was of the opinion that the extreme forms were genetically distinct. Weidenreich expresses himself positively against this opinion, justly, it seems to me. In fact, the mesamoeboids in young embryos vary much in size (compare Fig. 354) and a similar unevenness prevails also among the ichthyoid cells (Fig. 357). It is further probable that the large mesamoeboids, of which the majority form small cells by continual division, in small part at least develop haemoglobin precociously and thus produce the so-called megaloblasts.

Variation of the erythrocytes is especially pronounced in quite young embryos (Kolliker, 1846) and diminishes rapidly with age. At the close of fetal life it is comparatively slight. A statistical investigation of the variations in man is much to be desired.

An observation which I have occasionally made may be interpolated here. Now and again one finds a human embryo in which the erythrocytes contain from one to three small rounded granules, which are yellowish brown and vary in size and form. They are highly refractile and have no resemblance to nuclear fragments.

18 G. Bizzozero : Ueber die Entstehung der rothen Blutkb'rperchen wahrend des extra-uterinen Lebens. Moleschott's Untersuchungen zur Naturlehre, vol. xiii, 1888, p. 153-173, 1 pi.


These cells are especially numerous in an embryo of 6 mm. (No. 241 of Professor Mall's collection). Their significance is unknown to me.

Under pathological conditions granules may occur in the cytoplasm of erythrocytes which differ both from nuclear fragments and from the granules just described. They are unevenly fine granules, which take a basophile color. Naegeli 16 asserts that similar granules occur normally in the embryonic erythrocytes of several mammals (and also of man). I have been unable to confirm his statements.

The sauroid blood-cell changes into a blood-plastid by the loss of its nucleus. Since the change is imperfectly known in man, the following description is applicable rather to mammals in general than specifically to Homo. According to the original view of Kolliker " the nucleus was dissolved within the cell. According to the view of Rindfleisch" the nucleus is expelled. Both views have had many subsequent defenders. Jolly, I.e., gives an excellent exposition of the whole discussion. A more condensed review is given by van der Stricht. 19 Jolly reaches the conclusion that karyolysis may occasionally occur in young embryos, and is very

Fig. 358. — Four blood-corpuscles from a human embryo of 15.5 mm. X 1500. Coll. F. P. Mall, No. 390.

Li., nucleus of liver-cell for comparison of sizes.

rare, or does not occur at all, in older embyros and after birth. On the other hand, the expulsion of the nucleus is to be regarded as a normal process. The erythrocyte does not usually expel the whole nucleus at once, but in the form of single pieces which are driven out in succession. 20 Maximow, 21 however, reports that in young rabbit embryos the nucleus is expelled while still intact. It may happen that the part of the nucleus first expelled is larger than the part left behind. The expulsion may be easily observed in various phases; the phenomenon does not begin until "Naegeli: Ueber basophile Granulationen der Erythrozyten bei Embryonen. Folia haematol., vol. v, 1908, p. 525.

17 Kolliker: Ueber die Blutkorperchen eines menschlichen Embryo. Zeitschr. f. rationelle Med., vol. iv, 1846, p. 112.

"Rindfleisch: Ueber Knochenmark und Blutbildung. Arch. f. mikr. Anat., vol. xvii, 1880, p. 21.

" O. van der Stricht : Archives de Biol., vol. xii, 1892, p. 247. 'Man vergleiche Kostanecki, Anat. Hefte, vol. v, 1891, p. 315 and 317. A. Maximow: Arch. f. mikr. Anat., vol. lxxiii, 1909, p. 486.


DEVELOPMENT OF THE BLOOD. 500 the formation of blood has commenced in the liver. It occurs abundantly later in the bone medulla, but is rare in the blood-vessels of other parts of the body. Occasionally a pyknotic nucleus forms buds, which lead to the fragmentation of the nucleus and prepare for the expulsion. The expelled nuclei and nuclear fragments are, for the most part, eaten by phagocytes s and therefore almost never appear in the circulating blood. In passing, it may be mentioned that according to Afanassiew 23 the expelled nucleus becomes a blood-plate, an assumption which it has not been possible to affirm.

In young human embryos the blood plastids vary greatly; on the average they are larger than in later stages or in the adult. Maximow, 24 studying the rabbit, distinguishes the first erythrocytes as "primitive erythroblasts " and emphasizes the differences in their structure and that of later forms.

The early human plastids do not have the characteristic form of the definitive corpuscles, but retain a spherical shape. Gradually the cells of this type disappear, and at the same time appear



Fig. 359. — Blood-corpuscles from a blood-vessel of a human embryo of eight months. X 1500.

the small cup- shaped corpuscles (Fig. 359) which increase steadily in number. Meanwhile the nucleated erythrocytes gradually disappear from the blood, so that a little time after birth only the cupshaped corpuscles are found in circulation.

We can observe, very earlv, disintegration of the erythrocytes, 7 • %i 7 C_J * */ 7 even of the primitive mesamceboids. Three sorts of disintegration are to be considered: 1, dissolving of the haemoglobin and bursting of the corpuscle ; 2, fragmentation ; 3, vacuolization, with subsequent plasmolysis. Cells of the blood may die off in the most various stages of development. Their cytomorphosis closes with death. How far the death phenomena differ in the two cases is unknown. If the haemoglobin is dissolved out, the ervthrocvte remains as a round vesicle with or without a nucleus, as the case may be, with otherwise colorless contents and with a distinct

1 0. van der Stricht : Arch, de Biol., vol. xii, 1892, p. 251. Afanassiew: Deutsch. Arch. f. klin. Med., 1884, p. 217. 'A. Maximow: Arch. f. mikr. Anat., vol. Ixxiii. 1009. p. 471.

510 membrane. In the case of a plastid, the corpuscle swells by imbibition and assumes a round form. Although erythrocytes which have lost their haemoglobin occur frequently in embryonic blood, yet, as might be expected, it is very rare to get sight of one in the moment of bursting. The fragmentation of the red corpuscles in the adult has long been known. It occurs also in fetal life, but has as yet been little studied in embryos. The disintegration by vacuolization has, so far as known to me, not been described hitherto, 25 and consequently may be treated somewhat more fully. So far as my observations go, this form of disintegration occurs only outside of the vessels.

Any embryologist can easily convince himself that all forms of blood-cells occur in the mesenchyma of young embryos. I have


t >


. i\

Fig. 360. — An erythrocyte lying free in the mesenchyma. > 1800.

observed this distribution not only in man, but also in the pig, sheep, rabbit, cat, etc. There are no exceptions. Generally speaking, the forms of the corpuscles in the mesenchyma are identical with those in the vessels of the same embryo. Fig. 360 represents an erythrocyte lying free in the mesenchyma in the neighborhood of the forebrain of a human embryo of 6 mm. The red cells are widely scattered, but occur most frequently near the vessels. Sometimes they lie singly, and sometimes there are several together. The primitive colorless cells show a similar distribution in the mesenchyma. They are the so-called "primitive wandering cells" to which attention has been directed by several

28 We repeatedly find in the literature mention of wandering cells with vacuolated protoplasm, but they seem not to have been recognized as degenerating cells.

DEVELOPMENT OF THE BLOOD. 511 investigators, and especially by Saxer. 26 In my opinion, the conditions can be interpreted only on the assumption that all the free cells have wandered from the blood-vessels into the mesenchyma. 1 recognize no basis for assuming that we have to do with a progressive development in the mesenchyma, but this assertion is not equivalent to an absolute denial of such a possibility. It must be added that I have not yet been able to find evidence of the metamorphosis of mesenchymal cells into wandering cells. This metamorphosis has been especially emphasized by Maximow 27 and others, as a main part of their theories of the development of blood.

In older uninjured embryos we find that there are still mesamceboids, by no erythrocytes. How do the latter disappear? "We can answer that in part, at least, by degenerative vacuolization (compare below). The majority, according to an hypothesis I have formed, are removed by the lymph-vessels. This hypothesis is merely the application to mammals of a discovery made by Eliot R. Clark. Clark 28 observed, in living tadpoles, that erythrocytes which had passed out from the blood-vessels were overgrown by sprouts developing from the lymph-vessels, and thus brought into the cavity of the vessel, in which they then moved along centripetally. In support of this hypothesis may be mentioned the fact that in the placental chorion of man — which, as is well known, has no lymph-vessels — erythrocytes occur in the connective tissue up to the time of birth.

It has long been known that strikingly large free cells appear in the mesenchyma of the chorion. They are pictured in my ' ' Human Embryology. ' ' 29 Hofbauer 30 has recently again called attention to them. Grosser 31 mentions these cells — " deren Bedeutung aber noch unklar ist. Renewed investigation has led me to the conclusion that we have to do with erythrocytes which have gotten into the mesenchyma and, remaining there, have swollen by imbibition and are undergoing degeneration by vacuolization of their protoplasm. Fig. 361 represents eight of the cells referred to, from the chorion of an embryo of 15 mm. a is an unquestionable erythrocyte, although it exceeds somewhat in diameter the average red cells in other vessels, b is also an erythrocyte, but distinctly larger. We can explain the appearance of these cells by the assumption of imbibition, in which the nucleus has


Saxer : Anat. Hef te, vol. vi, 1896, p. 347.

A. Maximow : Arch, f . mikr. Anat., vol. lxxiii, 1909, p. 502. 28 E. R. Clark: Association of American Anatomists, Baltimore, 1909. See Anatomical Record, vol. iii, 1909, p. 183.

28 Minot : Human Embryology, Fig. 190, p. 330. 30 Hofbauer: Die menschliche Placenta, 1907.

Grosser: Eihaute und Placenta. Wien, 1900. p. 224.


,512 participated. Cells similar to a and b are easily found, but the majority of the cells in the mesenchyma have the habitus of d and e and exhibit the beginning of vacuolization. /, g, h are three cells which exhibit three stages of disintegration of the protoplasm. Since I have found similar cells in a considerable number of placentas, I draw the conclusion that they are constant and normal. I regard the interpretation of the pictures unattackable as proof of progressive degeneration. The cells g and h deserve special attention, because they look almost as though they were furnished with pseudopodia.

Now, we find similar denegerative appearances when we study the erythrocytes in the mesenchyma of the embryo. Hence we cannot avoid the conclusion that the corpuscles which have immigrated into the embryonic mesenchyma are subject to autolysis. I




Fig. 361. — Red blood-cells from the placental chorion of a human embryo of 15 mm. Coll. F. P. Mall, No. 350. a, from a blood-vessel; b-h, from the mesenchyma. X 1500.

believe that I recognize among the degenerating cells the so-called pseudopod-bearing cells which Maximow and others have described. Important, also, is the observation that the mesamceboid cells may degenerate in a similar way in the mesenchyma.

5. Hypotheses concerning the Formation of Erythrocytes. 82 — According to Hay em, 33 the red cells are developed from blood-plates, which he further calls haeinatoblasts. Pouchet (1879), Arndt (1SS1), Poljakoff, and others have defended Hayem's hypothesis.

The intracellular origin of blood-plastids has been asserted by Ranvier, 34 32 A much more extended analysis of the subject is presented by Jolly. Several older and often amusing hypotheses are mentioned by Feuerstack, Zeitschr. f. wiss. Zool., vol. xxxviii, 1S83, p. 136.

83 Das Hauptwerk Hayem's Du Sang erschein 1889. Darin stellte er die Ergebissne seiner fruheren Untersuchungen zusammen.

"Ranvier: Du Developpement et de l'accroissement des vaisseaux sanguins, Arch, de Physiol., vol. vi, 1874, p. 429-450.

DEVELOPMENT OF THE BLOOD. 513 Schafei', 35 Minot, and others. This conclusion was drawn from the observation of degenerating capillaries, which retain in their cavities blood-corpuscles and fragments of corpuscles after they have lost their connection with the active vessels. Ranvier named the degenerating capillaries " cellules vasoformatives." Vosmaer (1898) 3a made the true nature of these structures clear by his investigation of the embryonic great omentum. His discovery has been continued and extended by Renaut (1901), Pardi (1905), Jolly (1906), and others.

According to several hypotheses, the blood-plastids arise from the nuclei alone. Retterer 37 thinks they arise from the nuclei of connective-tissue cells. According to Hubrecht (1899), blood-corpuscles are formed in the placenta of Tarsius by the production of a colored mother cell which expels its nucleus, the nucleus becoming a red corpuscle. According to Poljakoff (1901) red disks are formed from the nuclei not only of connective-tissue cells, but also of leucocytes.

Since Neumann discovered (1869) nucleated red cells in the medulla of bones, there have been numerous hypotheses as to the method by which they changed into plastids. Many investigators have sought to recognize remnants of nuclei, or even entire nuclei, in the corpuscles after their transformation. In most of these cases one has to deal with artefacts. 38 Malassez (1881, 1882) lets the plastids arise as buds from the cytoplasm of red cells. According to Engel (1899) the red cell divides itself into a nucleated and a non-nucleated part; the latter is the definitive corpuscle. Janosik's hypothesis resembles that of Malassez. That the nucleus normally disappears by intracellular karyolysis has been a common opinion. 3 * An especially divergent account of the blood development in the yellow medulla of bone is given by Fr. Freytag. 40 He thinks that special cords arise by the degeneration of fat-cells. Into these cords blood-cells wander and there degenerate, their nuclei undergoing repeated fragmentation until they are broken up into small particles. The particles divide further until they become invisible, and these invisible remains of the nuclei he calls " nuclear units." These units gather to form new granules, the granules flow together and form genuine new nuclei, and finally new protoplasm is formed around each nucleus. The final member of the series is a new erythroblast. The author does not state how he has been able to follow the history of his invisible units, and does not show how he distinguishes the phases of evolution from those of involution of the blood-cells. Even if we admit the accuracy of these observations, they would still remain, in my opinion, without demonstrative value for the author's conclusions.

Our list of the hypotheses on the formation of the blood might easily be lengthened. Since, however, the hypotheses for the most part have only a passing interest, it is hardly worth while to go into greater detail. The reader will find further information given by Jolly. 41 6. Cttomorphosis of Leucocytes. — The primitive mesamceboids (primary wandering cells of Saxer, Maximow, and others) are also parent cells of the leucocytes, according to the conclusion drawn by Jolly and Weidenreich. Both of these authors have not only studied the literature conscientiously, but have also 85 Schafer : Note on the Intracellular Development of Blood-corpuscles in Mammals, Proc. Royal Soc, vol. xxii, 1874, p. 243-245.

88 Vosmaer: On the Retrograde Development of the Blood-vessels, etc.. Versl. Akad. Wetensk. Amsterdam, vol. vi, 1898, p. 245.

87 C. R. Retterer: Soc. Biol. Paris, 1901, p. 769.

    • Compare the careful discussion of Jolly, I.e., p. 180-193.

89 See especially Pappenheim, Virchow's Arch., vol. cxlv, 1896, p. 587. and vol. cli, 1898, p. 89; also His's Archiv, 1899, p. 214.

40 Fr. Freytag: Zeitschr. f. allgein. Physiol, vol. vii. 1908. p. 131. 41 Jolly, I.e., p. 180-193. An excellent, clear, conscientious review of the literature on the subject. Vol. II.— 33


made extensive independent investigations. While I here adopt their conclusion, I must admit that I cannot venture to express a secure judgment in this question, based upon my own experience. In a meritorious memoir Saxer 42 (1896) appeared as a defender of the view that free wandering cells (leucoblasts) arise directly from mesenchymal cells. Since then several authors have expressed their agreement with this view : for example, T. H. Bryee ** in his investigation of the development of the blood of Lepidosiren; and, recently, Maximow 44 in several articles, and also Weidenreich. 15 I have not succeeded in finding cell forms which can be unquestionably interpreted in favor of Maximow's opinion, although I have searched in numerous human and other mammalian embryos; and I must admit that the proofs which Maximow presents do not appear to me convincing. 4 " Therefore I keep, at least for the present, to the conviction that all leucocytes have a unitary origin and develop from the primitive mesamceboids.

A series of authors have defended the thesis that the first true leucocytes develop in the thymus from entodermal cells. Maurer 47 has maintained this thesis for teleosts, and it has been asserted for man and other mammals by Hermann and Tourneux, 48 Prenant, 49 , E. T. Bell, 60 and others. John Beard" has appeared as a specially eager defender. Stohr 52 opposes the thesis. Bryce, 83 Stohr, and Hammar 84 report that the true leucocytes first appear outside of the organ, and only secondarily, by immigration, in the thymus. The small cells which really develop in the thymus are derived, according to Stohr, from the epithelial cells and remain epithelial cells, not being lymphoid elements (leucocytes) at all. The question is of fundamental significance, although a priori it is improbable that leucocytes have a double origin. For the present I am much inclined to agree with Stohr, and we are thus brought back to the statement at the beginning of this section — the leucocytes are derived from the primitive mesamceboids.

The transformation is manifested by two principal alterations in the microscopic picture, — 1, the formation of special granules in the cytoplasm; 2, modifications in the form and structure of the nuclei.

We have to consider four principal types of leucocytes: 1. The young forms without granules (lymphocytes).

42 Fr. Saxer: Ueber die Entwickelung und den Bau der normalen Lymphdriisen und die Entstehung der rothen und weissen Blutkorperchen, Anat. Hefte, vol. vi, 1896, p. 347-532, Taf. xv-xxii.

48 T. H. Bryce: Histology of the Blood of the Larva of Lepidosiren, etc., Trans. R. Soc. Edinburgh, vol. xli, 1904, p. 435-467.

44 Maximow : l.s.c.

48 Fr. Weidenreich : Arch, f . mikr. Anat., vol. Ixxiii, 1909, p. 849-851, 857-858.

"I have been able to convince myself that Maximow is a very trustworthy observer, for I have confirmed many of his new observations by comparison with the sections in the extensive embryological collection of the Harvard Medical School.

47 Maurer: Schilddruse und Thymus der Teleostier, Morph. Jahrb., vol. xi, 1886, p. 129.

48 Hermann et Tourneux : Diet, encycl. sci. med., 1887.

48 Prenant : La Cellule, vol. x, 1894, p. 87-184.

60 E. T. Bell: The Development of the Thymus, Amer. Journ. Anat., vol. v, 1905, p. 29.

81 John Beard : Anat. Anz., vol. ix, 1894, p. 476-486 ; also Lancet, 1899.

'"Philipp Stohr: Ueber die Natur der Thymuselemente, Anat. Hefte, vol. xxxi, 1906, p. 407.

    • J. A. Hammar : His' Arch. Anat., vol. Ixxxiii, 1907, p. 83.

84 J. A. Hammar : Zur Kenntniss der Teleostierthymus, Arch. mikr. Anat.. vol. Ixxiii, 1908, p. 1-68, Taf. i-iii.

DEVELOPMENT OF THE BLOOD. 515 2. The older forms with granules — A. The finely granular (neutrophile of Ehrlieh).

B. The coarsely granular (eosinophile of Ehrlieh).

C. The degenerating (basophile of Ehrlieh).

1. The young forms probably arise directly from the primitive mesamoeboids, which have become smaller by repeated divisions. This origin of the lymphocytes was positively asserted in 1891 by 0. van der Stricht 65 and Kostanecki, 68 and is now very generally accepted. The number of leucocytes is also increased by their own proliferation. The lymphocytes vary extremely as to size. The large cells are probably (1) genuine primitive mesamoeboids, which by division produce the small leucocytes; (2) old cells, which have developed out of the small ones. 87 The following description is restricted to the small leucocytes, i.e., to the cells to which exclusively Ehrlieh M applies the term lymphocyte. Our cells have the following characteristics: first, they have very little protoplasm, which takes a basic color and exhibits no special granules; second, the colorable substance of the nucleus forms several little masses, often with distinct corners, which are united by threads and lie, for the most part, near the surface. The centrosome in the lymphocytes of amphibia has been studied by Flemming, Heidenhain, and Klemensiewicz. Weidenreieh (Arch. f. mikr. Anat., vol. lxxiii, p. 818) found the human centres double and imbedded in a lighter colored oval court situated in the endoplasm close to the nucleus.

Fig. 362. — Four small lymphocytes from normal human blood. (After Weidenreieh.) The developmental history of these cells is still incompletely known to us. That they multiply by mitotic division of the lymph-glands was first demonstrated by W. Flemming. 68 Whence they come and how they or their mother cells get into the lymph-glands is still to be determined. It is also unknown how the highly characteristic nucleus is formed.

That the lymphocytes are preliminary stages of the granular leucocytes is positively asserted by Weidenreieh. 80 As it is probable that he is right, lymphocytes

66 0. Van der Stricht : Le developpement du sang dans le foie embryonaire, Arch, de Biol., vol. xi, 1891, p. 19-113.

68 K. von Kostanecki: Anat. Hefte, vol. i, 1892, p. 313.

61 As concerns the nomenclature, see Weidenreieh, Arch f. mikr. Anat., vol. lxxiii, 1909, p. 794.

™ Ehrlich's application of the term " lymphocyt " in this restricted sense cannot be justified. Compare Weidenreieh, Arch. f. mikr. Anat., vol. lxxiii, p. 797 ff.

"W. Flemming: Arch. f. mikr. Anat., vol. xxiv, 1885, p. 50.

80 And by others before him. Compare W. H. Howell, Journ. of Morph., vol. iv, p. 144; C. Benda, Arch. Anat. Physiol., physiol. Abth., 1896, p. 347; and Weidenreieh, Arch. f. mikr. Anat., vol. lxxiii, 1909, p. 861.


are here regarded as the representatives of the young stage in the cytoruorphosis of white blood-corpuscles. Unfortunately, the further development of these " young " cells is hardly better known to us than their origin.

The cells which have been recognized with certainty as becoming granular leucocytes are distinctly larger than the lymphocytes. If, therefore, they develop from the lymphocytes, we must say that during the process the protoplasm and the nucleus have both grown. The protoplasm retains its capacity of basic coloration; the nucleus retains — at least at first — its round form, has in its interior a coarse reticulum with some few thicknesses, and it stains deeply. Very often the centrosome can be seen in an eccentric position alongside the nucleus, and its occurrence is probably constant. Cells of this kind occur throughout life in the medulla of bone, and are well known to histologists under the inappropriate name "myelocytes." Out of such cells the three kinds of granular leucocytes are developed.

2, A. The finely granular leucocytes are much more numerous than the coarsely granular, and they represent the chief developmental series of the white corpuscles. The granules in man are " neutrophile," in the rabbit " pseudoeosinophile," and in the guinea-pig " aruphophile." According to Ehrlich, the coloration of the granules changes with the age or maturity of the cells. The clinicians, for diagnostic purposes, have occupied themselves industriously with the question of the coloration of granules and have founded a doctrine of the specific quality of the granules of leucocytes based on the coloration. Up to the present, however, the proof is entirely lacking that the coloration in this case has morphological meaning, or even that it allows a deduction as to the chemical specificity of the granules.' 1 The alterations in the nuclei during the cytomorphosis of the finely granular leucocytes are very striking. The alterations begin with an elongation of the nucleus (Fig. 363), which, however, remains a unitary structure while assuming a kidney-like shape and at the same time moving into a decidedly eccentric position. The convexity of the nucleus is directed to the exterior; the concavity is turned toward the centre of the cell. In the central part of the cell lies the centrosome. By the deepening of its concavity, the nucleus becomes sausage or horse-shoe shaped and at the same time grows more slender and longer, so that the two poles of the nucleus move toward the non-nucleated side of the cell and approach one another. In the next stage the nucleus appears divided into several small pieces, which are connected by thin short or long threads. The number of pieces is usually three or four, seldom five. The form and size of the pieces is extraordinarily variable. Uniting threads may start from any point on the surface of the pieces. During this transformation of form the nucleus elongates and becomes more bent, always curving around the centrosome. The nucleus further undergoes a pronounced pyknosis, so that when it reaches the lobate condition it exhibits no recognizable structure, but stains intensely and more or less uniformly. The alterations in shape are permanent and are not transitory consequences of amoeboid movement of the nucleus. We have to deal with a genuine cytomorphosis : the nucleus never 81 Compare Fr. Weidenreich : Arch, f . mikr. Anat, vol. lxxii, 1908, p. 308-319.

DEVELOPMENT OP THE BLOOD. 517 turns back in its course of development. The centrosome 62 has usually two centrioles which are round or oval and usually of the same size. A single centriole occurs rarely, and probably arises by the fusion of the two normal centrioles. The centrioles are surrounded by a small clear court of apparently homogeneous material, but which sometimes shows a radiate structure. When the disintegration of the cells begins, the centrosome can no longer be seen.

The formation of the fine granules in the protoplasm may begin either as soon as the nucleus assumes its eccentric position or not until it has reached the lobate form. The granules are small, more or less uniform, and of apparently round shape, and arise endogenously. The general attention of hsematologists and clinicians has been directed to them by the invention by Ehrlich of a method of demonstrating them easily. They appear first at one or several points in the cytoplasm, and increase gradually in number until they occupy the whole body of the cells, with the

• •-

Fig. 363. — Finely granular (neutrophile) leucocyte with compact nucleus, a so-called myelocyte. From the circulating blood of a healthy adult. (After Weidenreich.) exception of the immediate neighborhood of the centrosome. Meanwhile the colorability of the intergranular substance diminishes. In a human embryo of three and one-half months the majority of "leucoblasts" in the bone medulla are without granules, but both neutrophile and eosinophile cells are present. During the fourth and fifth months the finely granular cells become more numerous.

2, B. The coarsely granular leucocytes (eosinophiles) develop at the same time as the finely granular, but are morphologically wholly different. The original round nucleus becomes eccentric and kidney-shaped, then more slender and longer, and bends around the centrosome, which takes a central position. Up to this point of its development it can scarcely be distinguished from the nucleus of the finely granular cell, but finally it assumes its permanent shape by forming two lobes (Fig. 364), which are united

82 The fine research of M. Heidenhain upon the leucocytes of Salamandra (Festschr. f. Kolliker, 1892, p. 138) must be regarded as the starting-point of our knowledge of the centrosome of leucocytes.


by a strand varying in width and length. As a rule, the lobes are of unequal size, but the inequality is seldom striking. The lobes are in general round or oval and occasionally pear-shaped. They usually have very regular contours, but occasionally one sees a small projecting hump. The length of the uniting strand is very variable. Nuclei occur with shapes which do not exactly fit with this description, but they are extremely rare. The centrioles are similar to those of the finely granular leucocytes, and are also situated in a clear court of material, which, however, in the case of the coarsely granular cells, is surrounded by a darker, broader zone, which appears nearly homogeneous. The darker zone is often extremely distinct.

The eosinophile granules, according to Weidenreich, 63 are not endogenous structures, but are fragments of erythrocytes which have been eaten by the cells. He found that the erythrocytes break

364. — A coarsely granular leucocyte (eosinophile) with a bilobate nucleus. Highly magnified. From the blood of a healthy adult. (After Weidenreich.) up into fragments in the haemolymph glands of sheep and other mammals, and that the fragments break up into still finer granules, which retain their characteristic color reaction, and are taken up by the lymphocytes, in which they appear as the eosinophile granules. In the measure that the number of granules in the single cells increases, the nucleus passes through its metamorphosis. In the finely granular leucocytes, on the contrary, the granules arise sometimes earlier, sometimes later. A renewed investigation of the eosinophiles in man is very desirable.

2, C. The degenerating leucocytes correspond to the " basophils" of Ehrlich's nomenclature. They are designated by Maxi 63 Fr. Weidenreich: Anat. Anzeiger, vol. xx, 1902, p. 196; also Verhandl. Anat. Ges., vol. xix, 1905, p. 79, and Arch. f. mikr. Anat., vol. lxxxii, p. 282 and 286 (extended discussion). The explanation adopted hy Weidenreich had been previously proposed by Hoyer, Klein (Cbl. inn. Med., 1899), and Fuchs (ditto). Weidenreich's observations have been confirmed by Warthin and Th. Lewis. Zietschmann joins in the opinion of these authors.

DEVELOPMENT OF THE BLOOD. 519 mow 64 as "Mastleucocyten." 65 An apposite name for these cells is still lacking. As Maximow has demonstrated, we must distinguish strictly, morphologically and genetically, between "Mastzellen" and " Mastleucocyten." The Mastleucocyten make up a very small percentage of the leucocytes in normal human blood, but occur abundantly in many cases of pathological blood. The nucleus becomes first kidney-shaped and then of an irregular contour, as may be seen in Fig. 365. As the alteration continues, pieces of the nucleus pinch themselves off from the main mass, or else the nucleus assumes a highly irregular form and breaks up into single pieces. These alterations may be easily observed in leukaemic blood. No distinct internal structure can be recognized in the nuclei. The amount of chromatin must be increased, for the coloration of the nucleus is more intense than in other leucocytes. There is no centrosome. The granules are extremely variable in number, size, and form ; in some cells there are only a few present,

Fig. 365. — Degenerating human leucocytes (Mastleucocyten of Maximow). (After Weidenreich.) in others they are abundant. The size varies also within a single cell. The form of the granules is often strikingly irregular; it may be angular, rounded, or elongated, or, in another case, the granules may be more uniformly rounded. They stain a dark blue violet with the Griemsa solution. The protoplasm loses its basophile reaction and appears vacuolated, especially in cells the nuclei of which have become irregular.

Note. — So far as known to me, Blumenthal " was the first to interpret these cells as degenerative, a view to which Pappenheim ** and Weidenreich w have agreed. That this interpretation is correct is rendered probable by the above-given history of the cells.

"Maximow: Ueber die 3ellforaien des lockeren Bindegewebes, Arch. f. mikr. Anat., vol. lxvii, 1906, p. 706.

65 Die " Mastleucocyten " des Meerschweinchens und anderer Rodenten sind morphologische vollkommen verschieden von der hier zu beriicksichtigenden menschMcJien Zellen.

98 R. Blumenthal : Ann. Soc. R. Sci. Med. et Nat. Bruxelles, vol. xiv, 1905.

47 Pappenheim : Atlas der mensehlichen Blutzellen, 1 Lief., 1905.

88 Fr. Weidenreich: Folia haematologica, vol. v, 1908, p. 135; also Arch. f. mikr. Anat., vol. lxxii, 1908, p. 252.


The multiplication of leucocytes is effected, as above stated, p. 515, by mitotic division, 69 which may often be seen in the young forms. As the cytomorphosis progresses, the mitoses become rarer. Blumenthal 70 and Benaut 71 have shown that the mitoses continue in the finely granular and eosinophile cells up to the stage of the kidney-shaped nucleus. In cells with lobate nuclei, mitoses have not been observed.

Amitosis of leucocytes has been described by several investigators. H. Pollitzer 72 has given a compilation of the recorded statements. Since the process has been observed only in nuclei which have become pyknotic, it is probable that we have to deal with a degenerative process which accompanies the downfall of the cells and plays no role in their normal multiplication. Yet authorities are not lacking to defend the hypothesis that the normal multiplication of leucocytes is by amitosis. Lowit 73 goes very far in this direction, for he asserts that the multiplication of leucocytes in the liver is effected only by amitosis, a view which Kostanecki 74 has shown to be completely untenable.

Disintegration of Leucocytes. — We may assume that leucocytes at the close of their cytomorphosis die and disappear. Death may befall a cell at any time, but accidental death is by no means comparable with the death which ensues at the close of the cytomorphosis of the cell. Our knowledge is still so incomplete that the following exposition can be regarded as tentative only. So far as the nuclei are concerned, we see that they break down by fragmentation, which begins with the rupture of the threads uniting the single lobes. Thus the cell becomes apparently multinucleated. After this the cells are often devoured by phagocytes ; but if they remain free, the splitting up of the nucleus continues until some ten or fifteen fragments are produced, which possess a rounded form and lie irregularly scattered in the protoplasm. The nuclear fragments, or lumps of chromatin, are homogeneous and disappear by dissolution. During the degeneration of the nucleus the granular cytoplasm gradually disappears. Finally, the remnant of the cell breaks down and the fragments are eaten by phagocytes, or perhaps in part dissolved.

68 Discovered by W. Fleming, Arch, f . mikr. Anat., vol. xxiv, 1885, p. 50-91. Compare also 0. van der Stricht, Verh. Anat. Ges. Gottingen, vol. vii, 1893, p. 81; and Jolly, Arch. d'Anat. rnicr., vol. iii, 1900, p. 168-228. 'Ihe latter gives a detailed analysis of the previous literature.

70 Blumenthal : Travaux Lab. physiol. Inst. Solway, vol. vi, 1904.

71 J. Renaut : Arch. d'Anat. micr., vol. ix, 1907, p. 495.

72 H. Pollitzer : Beitrage zur Morphologie und Biologie der neutrophilen Leueocyten, Zeitschr. Heilk. Abth. path. Anat., vol. xxviii, 1907, p. 277.

73 Lowit: Sitz.-ber. Akad. Wiss., Wien, vol. xcii, 1S85, 3 Abth. 74 Yon. Kostanecki, Anat. Hefte, vol. i, 1S92, p. 312.

DEVELOPMENT OF THE BLOOD. 521 The disintegration of leucocytes occurs in connective tissue, in exudates, and especially in the spleen and other lymphoid organs. Disintegration of the leucocytes also occurs in normal blood, but is rare.

7. Obiglx of the Blood-plates. 75 — We are indebted to the investigations of James H. Wright 76 for the recognition .of the actual development of the blood-plates. He succeeded in making the process clear by the application of a new method of coloration. 77 According to Wright, the plates arise by the pinching off of the ends of slender processes of uninucleate giant cells (the megakaryocytes of Howell). After the application of Wright's stain, one can see in both the blood-plates and in the giant cells a narrow hyaline-blue border (ectoplasm) the edge of which is either smooth or finely dentate. The breadth of the border varies, yet is the same in the two structures. Its peripheral layer is capable of amoeboid motion. The central portion of the blood-plate — as also the inner and by far larger portion of the cytoplasm of the giant cell — appears, after the same coloration, to be filled with more or less crowded granules of a red or violet tint. The majority of the giant cells — these observations refer chiefly to the bone medulla of various mammals — have a rounded form, while the minority exhibit forms of great diversity, which arise by the formation of pseudopod-like processes of variable length, width, and shape. In some cells almost the entire cytoplasm is absorbed in the formation of processes. We can observe, in these giant cells of a changed shape, that the red or violet granules of the internal substance of the cytoplasm extend into the processes and form in them an axial cord, which remains surrounded by a hyaline ectoplasm (Fig. 366). Occasionally a process extends into the cavitv of a blood-vessel. Some of them lose the connection with the parent cell, and such free pseudopods have been observed by Wright not only in the blood-vessels of the medulla of bone and in the spleen, but also in the capillaries of the lungs. Now in some of these processes, the width of which corresponds to the diameter of the blood-plates, we can see that the granular internal substance exhibits constrictions. At other points the subdivision of the middle substance is complete, and we encounter a series of rounded segments having the diameter and other characteristics of the internal substance of the blood-plates. Each segment of the internal substance has a clear peripheral zone. A process thus "Professor J. H. Wright has laid me under great obligations by reading the MS. of this section, and the value of the exposition has been much increased by his advice and additions.

76 James Homer Wright : Die Entstehung der Blutplattchen, Virchow's Arch., vol. elxxxvi, 1906, p. 55-63; see also Journ. Morphol., vol. xxi, 1910, p. 263.

77 Pathological Technique, by Mallory and Wright, 4th edition, 190S. p. 374.



modified is regarded by Wright as a chain of blood-plates which become free by the breaking np of the chain. Other processes occur which are so small that they probably produce only a single plate. That the giant cells really lose their protoplasm is proved by the occurrence of degenerating nuclei which are surrounded by little or no protoplasm.

Fig. 366. — Giant cells with processes from which blood-plates arise. Alongside are blood-plates and a few leucocytes. A, from the spleen of a kitten; B, from the bone-marrow of a cat. Original drawings by J. H. Wright.

Two further facts deserve especial attention in discussing "Wright's conclusions: first, that genuine blood-plates and genuine giant cells occur only in mammals; second, that the blood-plates first appear in the embryonic blood after the giant cells have been produced in the blood-forming organs.

Note. — A letter from Professor Wright enables me to add the following: There occur in the blood of mammalian embryos, before the development of blood in the liver has begun, together with a few blood-plates, a small number of cells which in their color reaction and in the structure of their protoplasm resemble the

DEVELOPMENT OF THE BLOOD. 523 giant cells of later stages, although in size they merely equal the red bloodcorpuscles. Wright has observed the cleavage of these cells into blood-plates. They occur in the embryos of guinea-pigs of 4.5 mm., but were not found in a younger embryo. By his investigations of other mammalian embryos, Wright has convinced himself that the cells in question, which occur free in the blood, are identical with the giant cells of the blood-forming organs. He has found all possible transitions. According to Maximow 78 the giant cells in the rabbit and other mammals arise from the primitive mesamceboids (primary wandering cells), which would agree with Wright's conclusion.

8. The Composition of the Blood in Eelation to Age. 79 — The distribution of the blood-corpuscles after the circulation has begun is probably always very unequal. This depends in part on the fact that the young blood-cells accumulate in special places, particularly those which serve as sites for the production of blood, concerning which see the following section. In the adult the percentage of the various forms of corpuscles differs according to the vessel. In embryos the relations are further complicated by the alterations which occur corresponding to the age.

The circulation of the blood begins extraordinarily early in man. The cytology of the blood at this moment is unknown to me.

In an embryo of 4 mm. I find large ichthyoid cells, as described on p. 505 and pictured in Fig. 355. Older cell forms are entirely lacking. Noteworthy is the extreme rarity of the primitive mesamceboids, a fact which does not correspond to my a priori expectations.

In an embryo of 7.5 mm. the cells are smaller on the average, but are still ichthyoid, Fig. 356, p. 506. They vary much in size, and may possess nuclei which indicate by their lessening diameter and deeper coloration the further cytomorphosis. In this case also I missed the primitive mesamceboids in the blood.

In embryos of 8-10 mm. the blood-cells are for the most part unquestionably ichthyoid, although their dimensions are extremely variable (Fig. 367). The younger types of cells are still extremely rare. One sees now and then accumulations of undifferentiated cells (Fig. 368) the protoplasm of which seems fused. Such clusters of cells adhere to the endothelium without being continuous with it.

Maximow 80 has observed similar clusters in the rabbit. According to his interpretation they arise by the proliferation of the endothelium. I am unable to

78 70

Alex. Maximow: Arch. f. mikr. Anat., vol. Ixxiii, 1909, p. 491.

The principal work on this subject is that of Johann Jost (Arch. f. mikr. Anat., vol. lxi, 1903, p. 668), but he investigated only sheep and cow embryos. On p. 691 he gives curves of the percentages of corpuscles. His Metrocyten I are ichthyoid cells, his Metrocyten II are sauroids, and his Erythrocyten red plastids. Valuable data concerning the relations in rabbit embryos have been published by A. Maximow (Arch. f. mikr. Anat., vol. Ixxiii, 1909, p. 526-532). 'Maximow: Arch. f. mikr. Anat., vol. Ixxiii, 1909, p. 517.

SO '

524 agree with hire, because I find that there is no continuity of the protoplasm of the cells either in the rabbit or in man ; also because mitoses of the endothelium in the neighborhood of the clusters are almost invariably lacking; and, finally, because the endothelial nuclei are differentiated while the nuclei of. the cells of the clusters are not differentiated. 81 The clusters may be compared with blood-islands, and I regard the cells composing them as mesamoeboids or primary wandering cells.

In all human embryos up to 12 mm. which I have had an opportunity of investigating, there occurs an active mitotic division of the red ichthyoid cells. Since the primitive cells are rare, we must conclude that the number of corpuscles increases chiefly by their own division during this period of development.

In embryos of about 12 mm. the blood formation is beginning in the liver, p. 528. At the same time the undifferentiated bloodcells become more numerous in the vessels, and the first cells of

Fig. 367. — Outlines of erythrocytes of a human embryo of 8 mm. Harvard Embryol. Coll. No. 817.

the sauroid type appear. From this time on, the sauroid cells become constantly more numerous. The red cells are very variable. It must be further remarked that probably many erythrocytes are destroyed in the blood itself, so that we encounter the following cell forms : First, red cells of the round shape with a distinct membrane, but without haemoglobin ; second, similar cells collapsed; third, nuclei with remnants of a cell body; and fourth, free nuclei. 82 I have not obsc wed expulsion of nuclei in very young embryos.

81 Compare Minot, " Age, Growth, and Death," Fig. 61, Nos. 5, 6, 7, 8. The nuclei of the cell clusters are all in the second stage in the figure referred to, and this stage immediately precedes the differentiation proper.

82 The possibility remains that in these cases we have to do, in part at least, with the consequences of imperfect preservation. Still, I eonsiJ ov it probable that the break-down occurs normally in the manner indicated in the living blood.



In embryos of two months 83 the blood contains, 1, a minority of ichthyoid cells; 2, a large majority of sauroid cells which may be easily recognized by their pyknotic nuclei; 3, a considerable number of non-nucleated plastids, the formation of which occurs chiefly in the liver ; 4, free erythrocyte nuclei, which are rare ; and 5, mesamceboid cells of various appearance, some larger, some smaller, the largest equal in diameter the nucleated erythrocytes.


Fig. 368. — Endothelium and blood-cells from the lower part of the aorta of a human embryo of 9.4 mm. Harvard Embryol. Coll., No. 380. Endo., endothelium; A., collection of cells; E., erythrocyte. X 1500.

Most of the mesamceboids retain their primitive character. Now and again, however, I found one with a kidney- shaped nucleus which probably was a mature granular leucocyte. Unquestionable "lymphocytes" I did not recognize.

83 1 have investigated seven embryos of about this age. In several the preservation of the tissues is pretty good, but in none is the preservation of the erythrocytes satisfactory. Therefore the statements given in the text possess only a preliminary value.


During the third month the young forms of the erythrocytes become steadily rarer and the ichthyoid cells almost disappear from the blood, while the blood-plastids become steadily more predominant. 84 In a beautifully preserved embryo 85 of about eight months, I have found the following conditions : By far the majority of the corpuscles are thin disks without nuclei, many of them shrunken; the unaltered disks are convex on one side and concave on the other. In Fig. 369 two such are figured in optical section. Nucleated erythrocytes are extremely rare. Free dark nuclei and occasionally fragments of nuclei appear now and again. The colorless cells form distinctly a minority, but may be found everywhere. They are either primitive mesamceboids or young leucocytes, rarely leucocytes with lobed nuclei. The lymphocytes have the above described structure of the nucleus (p. 515) which is so

Fig. 369. — Blood-corpuscles from the vessels of a human fetus of eight months.

highly characteristic. The eosinophile leucocytes are very rare; they have for the most part a round or oval nucleus. Only by searching can one find an eosinophile with a kidney-shaped nucleus. The coarsely granular leucocytes are more numerous in the thymus and lymph glands.

It remains for the future to furnish satisfactory data concerning the composition of fetal blood and the changes it undergoes corresponding to age.

The percentage relations of the white corpuscles after birth have been investigated by Carstanjen. 86 I have put his chief conclusions in the form of a table. According to Carstanjen, the number of coarsely granular cells does not depend on the age, and, as far as they are concerned, only individual variations were "For example, in an embryo of 29 mm. (Harvard Embryol. Coll., No. 914), although nucleated corpuscles are still numerous, the majority of the erythrocytes are without nuclei.

85 1 am indebted to Professor W. T. Councilman for this very beautiful material, preserved in Zenker's fluid, for which I here express my thanks.

"Carstanjen: Jahrb. Kinderheilk., 1900, p. 215 and 684.

DEVELOPMENT OF THE BLOOD. 527 observed. Very striking is the rapid increase of the young forms in the first days after birth. During the fifth year cells with lobate nuclei reach their maximum.

Percentage of Leucocytes in the Blood.

diately First Second

after Twelve half half Two Three Four Five birth. days.






years L6.0 45.6 50.8 49.2 47.0 38.4 33.2 25.1

Young forms (Lymphocytes) 16.0 Finely granular cells with lobate nuclei 73.4 36.7 34.5 40.8 42.0 48.0 52.6 61.0 9. Sites of Blood Fokmation. — Since it is highly probable that all blood-cells are descendants of the primitive mesamceboids, we must assume that these last seek out at different ages certain localities for their multiplication and transformation. 87 Of such sites of formation five are known with certainty : 1, the yolk-sack ; 2, the blood-vessels of young embryos; 3, the embryonic liver; 4, lymph-organs; and 5, the medulla of bone. 88 The spleen may occupy a special position among the lymphoid organs, since nucleated erythrocytes occur in it during the fetal period. According to Neumann, 89 the erythrocytes are simply swept in with the blood current, and are not formed in the organ itself. Investigators are by no means agreed, however, in their opinions concerning the actual process. It would be out of place here to enter on an extended discussion, therefore we give only a short account of the phenomena which is necessarily of a somewhat preliminary character. As has been explained above, I cannot accept the opinion that the general mesenchyma serves as a site for the formation of blood. This opinion has been recently defended — perhaps correctly — by Maximow. In this connection mention must be made of the interesting observations of Pardi, 90 who studied the blood-cells in the mesenchyma of the omentum in rabbits.

1. The Yolk-sack. 91 — That the earliest blood-corpuscles arise in the wall of the yolk-sack has long been known, and has been asserted for man (compare above, p. 501).

87 E. Neumann (Virchow's Arch., vol. cxix, p. 393) still defended, in 1890, the idea that the hsematoblasts arise in place in the medulla of bones, and that they are not immigrant elements.

™ The significance of the medulla of bone for blood formation was discovered by E. Neumann (Cbl. med. wiss., 1868), and almost at the same time by G. Bizzozero {ibid., 1869). Its discovery must be honored as the starting-point of the modern doctrine of blood development.

" E. Neumann: Arch. f. Heilk., vol. xv, 1874.

  • ° F. Pardi : Eritrociti nucleati, . . . nel grande omento nel coniglio, Arch. Ital. Anat. Embriol., vol. iv., p. 370-386, Tav. liii-liv (1905).

n The conditions in the yolk-sack have been admirably described by Maximow, with special reference to the rabbit. See Arch. f. mikr. Anat., vol. Ixxiii, 1909, p. 457 and 476.


Further exact observations upon the hematogenic activity of the yolk-sac during its development are still lacking, both for man and for mammals in general. We can report only that during young stages there is a striking excess of the early stages of the erythrocytes in the blood-vessels of the yolk-sack, and that mitoses of the blood-cells are frequent.

2. The Young Blood-vessels.- — As is well known, Remak 02 first discovered the multiplication of the blood-corpuscles in embryonic blood-vessels. We now know that this multiplication occurs by mitosis of both the mesamceboids and young erythrocytes. The mitosis in man can be easily observed in well-preserved material.

According to observations on mammals, the mitotic figures disappear from the circulating blood soon after the formation of blood in the liver is well started.

3. Blood Formation in the Liver. — That the liver serves as a site for blood formation in mammalian embryos, from soon after its first formation up to the end of fetal life, was first suspected by Prevost and Dumas 93 and later by Eeichert and E. H. Weber. 04 Kolliker, 95 however, in 1846, gave the first definite proof of this important phenomenon, - when he published his discovery that special cells occur in the fetal liver which change into erythrocytes. Although since then there have been many investigations 96 upon the fetal liver, the process of blood formation has not yet been completely cleared up. In these investigations there has been no lack of opinions which have later been recognized as untenable. Such, for example, is the opinion of Neumann, according to which the corpuscles arise endogenously ; or the opinion of Foa and Salvioli, who derived the erythrocytes from hepatic giant cells.

The blood formation in the human liver begins in embryos of about 12 mm. in length. 97 At this time the hepatic cylinders are well developed, but are separated from one another by broad sinusoids. The endothelium of the sinusoids clings everywhere 82 R. Remak : Ueber die Entstehung der Blutkorperchen, Med. Zeit., Ver. Heilk. Preussen, vol. x, 1841, p. 127. Compare also Canstatt, Jahresber., 1841, p. 17, and Remak's Untersuch. Entwicklungs Ges. Wirbelthiere, 1851, p. 22.

"Prevost et Dumas: Developpement du Cceur et formation du Sang, Ann. Sci. Nat., vol. iii, 1824, p. 96—107, p. iv (le foie sanguifactif, p. 105).

M E. H. Weber: Zeitschr. f. rat. Med., vol. iv, 1846.

85 Kolliker: Zeitschr. f. rat. Med., vol. iv, 1846.

89 The following investigations may be mentioned : Fahrner : De Globuli sanguinis, Turici, 1845. Neumann: Berlin, klin. Wochenschr., 1871, p. 58; and Arch. f. Heilk., vol. xv, 1874. Foa e Salvioli: Arch, delle Sci. Med., vol. iv, 1880. M. B. Schmidt: Ziegler's Beitr., vol. xi, 1892, p. 199. The most exact investigations are those of van der Stricht, Archives de Biol., vol. xi, p. 19-113 and vol. xii, p. 235; and Maximow, Arch. f. mikr. Anat., vol. lxxiii, 1909, p. 533-546. Both authors studied chiefly rabbits. Maximow gives a good review of previous results.

87 This is according to my own observations and the statement of Schridde, Verhandl. deutsch. pathol. Gesellsch., 1907, p. 364.

DEVELOPMENT OF THE BLOOD. 529 closely to the hepatic cylinders. The broad blood-channels are clearly bounded; they contain blood-cells of varying appearance, bnt no true leucocytes, only rnesamceboids and young erythrocytes (Fig. 370). Besides these there are also blood-cells so placed that they appear as part of the hepatic cylinders, and these last are the beginning of the development of the blood-producing centres in the liver.

From the stage of 12 mm. on, the number of the blood-cells which apparently are included in the liver cylinders increases rapidly. The blood-cells gather in little groups, which interrupt irregularly the hepatic cylinders. When colored sections are examined, these groups are conspicuous because the cells, in consequence of the progress of their development, possess nuclei of diminished size which stain very deeply. The nuclei of the livercells are much larger and more lightly colored, and their substance is not condensed, but forms a loose network. In earlier stages the sinusoids of the organ are separated from one another only by the cylinders consisting of liver-cells. In the region of the clusters ^ — ^ a c d b Fig. 370. — Blood-cells from a hepatic vessel of a human embryo of 11 mm. Coll. F. P. Mall, No. 353. o, c, d, primitive rnesamceboids; 6, erythrocyte. X 1500.

of erythrocytes, an examination of sections of the liver gives one the impression that the structure has remained essentially as before except that the hepatic cylinders now seem to consist in part of erythrocytes. Certainly the clusters of blood-cells lie outside the direct blood-channels through which the blood flows freely. It need hardly be said that the red cells are morphologically never parts of the hepatic cylinders.

The shape of the clusters of blood-cells is very variable. They may be sharply circumscribed, rounded, or elongated, and distinctly separated from one another; but quite as often they are extended into prolongations, by which the neighboring clusters may be united with one another, so that here and there we get a network with ample nodes. The size of the single clusters is very inconstant. In the third month it is not rare to find clusters in a section of which one can count fifty or more cells.

As regards the constitution of the cells in the blood clusters, we must distinguish the colorless cells from those colored with haemoglobin. M. B. Schmidt estimated the number of colorless and colored cells to be approximately equal in a human embryo of nine and one-half months, and in his opinion this proportion holds true Vol. II.— 34


for mature and nearly mature embryos. In earlier stages the colored corpuscles predominate.

The orginally colorless cells must be classed as mesamceboids, which must not be confused with true leucocytes. Nevertheless, they have been quite frequently loaded with this name. They can easily be distinguished from true leucocytes, although they are the parent cells of white corpuscles as well as of the red. Figure 369 represents some cells in the open blood-channels of the liver of an embryo of 11 mm. Such cells are numerous in the hepatic vessels

Fig. 371. — Hepatic cylinders of a human embryo of 11 mm. Coll. F. P. Mali, No. 353. The blood-cells with small nuclei lie apparently in the substance of the liver-cells, -which have large nuclei.

at this time, although rare in the vessels elsewhere. KostanecM 98 calls attention to the fact that the colorless cells lie usually, if not always, against the wall of the vessels. Mesamceboids with two nuclei alike in size are not rare.

The two chief forms of cells in the blood clusters are irregularly distributed. Often one sees in a single cluster nucleated elements, colored with varying intensity by haemoglobin, alongside of uncolored corpuscles. Now and again a group consists of only colorless corpuscles, which usually lie so closely pressed together that they flatten one another and appear like a mosaic pattern.

98 Von Kostanecki: Anat. Hefte, vol. i, 1891, p. 308.

DEVELOPMENT OF THE BLOOD. 531 Again, the boundaries between the cells disappear and the relatively large nuclei then lie closely crowded in a common protoplasmic mass. Such groups resemble those which occur in the blood-vessels of young embryos (Fig. 368). They are to be regarded as developing mesamceboids.

The colored cells differ much from one another. They vary within wide limits of size, in this respect contrasting with the nonnucleated plastids. There occur many cells which are so large that the nucleus alone equals the average volume of the plastids. Others, on the contrary, are less in diameter than a plastid. The content of haemoglobin must be independent of the size of the cell and the amount of protoplasm; it now appears in the somewhat saturated color, as in the plastids, and again as a just-recognizable shade of vellow. Every intermediate condition occurs between these extremes. The cells containing the least haemoglobin may display a light, granular protoplasm which, as the cells become more strongly colored, can no longer be seen. The optical homogeneity of the mature erythrocyte is generally known. The nuclei vary in their structure: first occur those of the ichthyoid type, which by their structure are clearly related to the nuclei of the colorless cells ; second, nuclei of the sauroid type, which are smaller, and the substance of which appears darker and more homogeneous, allowing only a few indistinct granules and threads to be recognized in it. Every possible transition between these two nuclear types may be observed. The small sauroid nuclei belong to the cells richest in haemoglobin. One may say that the higher the content of the cell in haemoglobin, the smaller is the nucleus and the more condensed its chromatin framework.

"We come now to the question of the relation of the abovedescribed cells to the open sinusoids on the one hand, and, on the other, to the hepatic cylinders. Two opposing views are to be considered, since some defend the opinion that the clusters all lie in the vessels or in diverticula of the vessels, while others assert that a part of the cells are extravascular in position. Thus, M. B. Schmidt (I.e., p. 203), von Kostanecki," and others assert definitely that the cell clusters are intravascular; 0. van der Stricht, 100 on the other hand, reports that in mammals the young blood-cells lie in part outside of the vessels, between the hepatic cells, and that in older embryos the cells are clearly embedded in the mesenchymatous cells. So far as my observations go they fully confirm van der Stricht, 101 and the excellent observations on rabbit embrvos


Von Kostanecki: Anat. Hefte, vol. i, 1891, p. 308. O. van der Stricht, Arch, de Biol., vol. xii. 1892, p. 241. 11 The conditions in the opossum embryo appear to me especially clear and unquestionable.


which Alexander Maximow 102 has recenly published appear to me to decide the question. Finally, and in favor of van der Stricht, since the vascular endothelium of young embryos allows the bloodcorpuscles to pass through, we need feel no surprise that a migration of blood-cells occurs also in the liver.

By the shaking of sections, Schmidt has removed a portion of the blood-cells, and has then observed spaces which were simply enlargements of the sinusoids and which were enclosed on both sides by the cords of liver-cells. The liver-cells are pressed back, especially from the wider cavities, and are thereby reduced sometimes to mere strips of protoplasm ; the endothelium is frequently still recognizable. In other cases, Schmidt saw the blood-cells situated between two liver-cells ; the latter appeared hollowed out and as if eaten by the blood-cells (lacunar corrosion of Neumann). At other times it appeared to him as if the small blood clusters were embedded in a single liver-cell. In both cases he was unable to see the endothelium. Strictly speaking, therefore, the observations of Schmidt agree better with the conclusions of van der Stricht than with his own view.

Origin of the Cells of the Blood Clusters. — According to the prevalent view, as above stated, the clusters arise by the accumulation of cells which circulate in the blood. A fact in favor of this view is that from the first all the forms of cells occur in the blood clusters which at that time can be found in the blood. If we had to do with cells which arose in loco, we should expect that only young cells would appear at first, which later would differentiate themselves. Further investigations are necessary to give a final decision as to the origin of the cells.

M. B. Schmidt 103 has drawn from his observations the conclusion that the proliferation of the endothelium produces new young erythrocytes, which then multiply farther by mitosis. To me his argument is by no means convincing.

Investigations up to the present render it clear that a multiplication of erythrocytes occurs in the embryonic liver during a long period, the end of which is after birth. The cytomorphosis of the red cells in the liver is essentially, or exactly, the same as elsewhere during early stages of the embryo, and also as in the medulla of bone. The clusters yield, at least in part, immature corpuscles which enter the circulating blood, so that we must assume that these complete their cytomorphosis in the blood itself.

0. van der Stricht (I.e.), von Kostanecld (I.e., p. 313), and others report that in the liver of mammals lymphocytes are formed by the metamorphosis of primitive colorless cells.


Maximow: Arch. f. mikr. Anat., vol. lxxiii. 1909, p. 538. M. B. Schmidt : Ziegler's Beitrage, vol. xi, 1892, p. 212, 219.

DEVELOPMENT OP THE BLOOD. 533 Neumann was the first to prove that the blood flowing out of the liver contains more young erythrocytes and more mesamoeboids than the inflowing blood. Later M. B. Schmidt found that in a nearly mature embryo the proportion of nucleated to non-nucleated cells was — In the portal vein 1 : 38, In the hepatic vein 1 : 25.

4. Formation of Leucocytes in the Lymphoid Organs. — It is well known that the lymphoid glands, tonsils, the thymus, etc., all serve for the multiplication of lymphocytes. It is easy to observe the proliferations of the young leucocytes in the corresponding embryonic organs of man, yet the life history of the lymphocytes during embryonic development is almost unknown.

5. Blood Formation in the Medulla of Bone. — The medulla of bone 104 is a vascular, mesenchymatous tissue, which originally is merely a reticulum of branching cells, with relatively wide endothelial blood-vessels. A part of the cells become osteoblasts. In the adult condition the structure remains essentially the same, although the mesenchvma forms connective-tissue fibrils, multinucleated giant cells, and, later, fat-cells; the latter vessels become arteries and veins. Certain investigators assert that the cavities of the blood-vessels are in direct open connection with the spaces of the mesenchyma, but, so far as I know, the conclusive proof of the correctness of this statement is lacking.

The history of the fetal medulla in man has not yet been investigated, probably because fresh material is indispensable for such investigations. Hence it is that we can merely sketch the history in outline, and about as follows : The medulla of bone arises late ontogenetically, since it does not appear until immediately before the commencement of ossification in each piece of cartilage, and therefore appears in different parts of the skeleton at different times.

Soon after its formation, there appear in it cells which we regard as primitive mesamceboids, and also young erythrocytes, young leucocytes, and young giant cells (megalokaryocytes) ; all three of which, according to the well-founded prevalent view, are developed from mesamceboids. The mesamoeboids are called, inappropriately, " myelocytes," although as compared with the original medulla (mesenchyma) they must be regarded as foreign elements. Gradually the number of " myelocytes," as well as young red and white cells, increases. The cytomorphosis progresses toward maturity, that is, until the erythrocytes have lost their nuclei and the leucocytes have acquired their granules. When mature, the corpuscles under normal conditions pass into the blood stream, by which they are carried off in order to participate in the circulation until they break down. Probably the lymphocytes also pass in small numbers into the blood stream. Only under pathological conditions do nucleated erythrocytes or so-called myelocytes pass out


C. M. Jackson : Zur Histologic und Histogenese des Knochenmarkes, Arch. f. Anat, 1904, p. 33-70.


in noticeable numbers into the circulating blood. This pathological condition has as yet been observed only after birth.

The number of blood-forming cells in the medulla increases slowly until birth, after which it mounts rapidly in the course of a few days. Incidentally occurs a diminution of the blood formation in the liver.

The distribution of the " myelocytes " and developing blood-corpuscles in the medulla has not been rendered clear by the existing investigations. The uncertainty is chiefly due to the fact that neither the course of the smaller vessels nor the structure of their walls is sufficiently known to us. The developing blood elements lie in part in the mesenchyma, in part in wide capillaries, — according to several investigations, made chiefly upon rabbits. 108 I may cite especially the work of 0. van der Stricht and of Brinckerhoff and Tyzzer. The young forms are in excess in the mesenchyma. The young forms include the colorless mesamceboids, the ichthyoid erythrocytes, and the granular leucocytes with round or kidney-shaped nuclei. The cells in the vessels often cling closely crowded to the vascular wall, almost as if they were glued to it and to one another. In these vascular accumulations the sauroid erythrocytes and the mature granular leucocytes predominate. Since the number of the red plastids outside the active blood stream is never very great, we conclude that the plastids leave the medulla soon after their development.


The earliest developmental processes of the heart, especially in so far as they concern the formation of the endothelium of the heart and vessels, are unknown in the human embryo, but probably one will not be far astray in assuming that the earliest anlage of the human heart is essentially similar to that of the mammalia. The earliest development of the heart is naturally associated with the first appearance of the vessels, but concerning this the following brief statement is all that is necessary here. According to the comprehensive investigations of Mollier, the preliminary

105 Robert Muir : On the Relations of Bone-marrow to Leucocyte Production and Leucocytosis, Journ. Path, and Bacterid., vol. vii, 1901, p. 161.

Muir and Drummond : On the Structure of Bone-marrow, etc., Journ. Anat. and Physiol., vol. xxviii, 1893, p. 125.

Rindfleisch: Arch. f. mikr. Anat., vol. xvii, 1880, p. 1-11 (describes the circulation).

W. H. Howell: On the Life-history of the Formed Elements of the Blood, Journ. of Morphol., vol. iv, 3890, p. 57.

E. Neumann : Ueber die Entwickelung rothen Blutkorperchen in neugebildetem Knochenmark, Virehow's Arch., vol. exix, 1890, p. 385-398.

Freiberg: Experimentelle Untersuchungen fiber die Regeneration der Blutkorperchen im Knochenmarke, Inaug. Diss., Dorpat, 1892.

Bizzozero: Cbl. med. Wiss., 1881, and Moleschott's Unters. z. Nat.-Lehre, vol. xiii, 1888.

O. van der Stricht : Archives der Biologie, vol. xii, 1892, p. 199.

Brinkerhoff and Tyzzer: On the Leucocytes of the Circulating Blood of the Rabbit, Journ. Med. Research, vol. vii, 1902, p. 173.

THE DEVELOPMENT OF THE HEART. 535 to the formation of the heart in all craniote vertebrates is the appearance of a number of cells between the endoderm and mesoderm, at first in the distal portion of the head. These elements, known as vascular cells, are discernible much earlier in the amniota than in the anamnia, and are recognizable in mammalian embryos with two or three primitive somites. From these vascular cells there develops, however, only the cardiac endothelium, the remaining constituents of the heart wall, the myocardium and epicardium, being derivatives of the visceral ccelomic wall. The first aggregation of the vascular cells of the heart is paired and produces a bulging of the visceral lamella into the wide pleuropericardial cavity. This bulging portion of the wall, which, as already stated, gives rise to the entire heart wall with the exception of the endothelium, has been named by Mollier the heart plate or cardiogenic plate. The topical relation of the paired heart anla^en to one another, that is to say, the time when they come into contact, depends on the configuration of the foregut. If this* is |pread out flat at the time of the appearance of the heart anlagen, these are widely separated from one another; if, however, there is an early closure of the fore-gut ventrally, as, judging from stages already known, is undoubtedly the case in human embryos, then the paired heart anlagen are very close together from the beginning and their fusion takes place early. In the Spee embryo G-le (Normentafel 1 No. 2, primitive somites not yet visible) some scattered vascular cells occur in the region of the paired heart anlage.

On the closure of the fore-gut ventrally the hitherto symmetrical pleuropericardial cavities come together anteriorly and fuse in this region, the median partition between them, the mesocardium anterius, disappearing, while the mesocardium posterius persists for some time longer. The closely approximated but not yet fused endothelial tubes are now surrounded by a continuous myo-epicardial mantle (Mollier). Finally the two endothelial tubes come into contact, their partition wall disappears, and the unpaired heart cylinder is formed from the paired heart tubes. This stage of the development of the heart occurs in the KromerPfannenstiel embryo, Klb (Normentafel No. 3, five to six primitive somites), a section of which, passing through the heart anlage, is shown in Fig. f 72. The space which is seen between the myo-epicardial mantle and the endothelial tube, and which is probably filled with fluid intra vitam, is perhaps somewhat enlarged in this embryo by the collapse of the endothelial tube during preservation. The fusion of the endothelial tubes is shown

  • By Normentafel is meant Kernel's Normentafel zur EntwicklungsgescmVhte des Menschen.


only in a few sections: cranially and caudally from the section figured one still sees the paired ends of the endothelial tnbes.

With the fusion of the. paired anlagen to form an unpaired cylinder there begins a new period in the development of the heart, during which two processes take place simultaneously, namely, (1) the elongation and consecutive bending of the heart cylinder and (2) the differentiation of the heart into its individual parts. There is no doubt that both processes are the result of the functional elaboration of a primarily straight and simply propulsatory portion of the vascular system, that is to say, of the heart anlage. In the description that follows a division of the process into stages will be made; in the first stage the development of the heart wall will be followed from the condition of a simple straight cylinder, to which stage it has now been traced, up to the time of the primary atrial division. The second stage will extend from the development of the primary atrial septum to the degeneration of the septum primum and the development of the septum secundum, the third to the complete division of the heart, and the fourth, finally, to the acquisition of its definitive form.

The development of the mammalian heart from the stage in which it is a simple cylinder to the completion of its development has been made known by the fundamental work of Born. The observations of this author were made principally on the heart of the rabbit, but were frequently extended also to the human heart. Born has modelled some stages of the latter and has pointed out the slight differences that obtain in the development of the two hearts. Recently Hochstetter has written a comparative embryology of the vertebrate heart for Hertwig's Handbuch. The succeeding account of the development of the human heart follows closely the work of Born, yet the endeavor has been made to complete as far as possible our knowledge of the development of the human heart.

The bending of the heart cylinder begins by the portion exactly midway between the two fixed ends being thrown into a loop, which may be termed the ventricular loop and whose apex is towards the right. The cranial end of the heart cylinder is fixed at the point of emergence of the cylinder from the pericardium, that is to say, at the point of division of the truncus arteriosus into the aortic arches; the venous end, to which the umbilical and omphalomesenteric veins converge, is fixed by the septum transversum, developing immediately above the yolk-sack. Since the heart cylinder grows more rapidly than the fixed points separate from one another, its free portion becomes thrown into a loop. The two limbs of this loop are separated by an almost horizontal cleft, the interventricular, or better the bulb o -ventricular cleft (compare Fig. 373 2 ). By this cleft the heart loop is sepa 2 1 am indebted to Professor P. Thompson for the loan of this model.



rated into two portions, into a cranial limb, the bulbar limb, and a caudal one, the ventricular. The part situated immediately above the septum transversum widens later to form a cavity whose greatest diameter is transverse and whose left end communicates with the ventricular limb. This cavity represents the atrial portion of the heart, and the somewhat constricted portion by which it communicates with the ventricular limb is the atrial canal. The atrial portion lies dorso-caudal to the ventricular limb, which, on its pari, is overlapped cranially by the bulbar limb. On the caudal wall of the atrium there opens the sinus venosus, which is greatly expanded in the transverse direction and whose cavity receives the blood from both venae umbilicales, venae omphalomesentericse, and ductus Cuvieri. Consequently, simultaneously with its bending,


Fig. 372.

Fig. 372. — Section through the heart anlage of the Pfannenstiel-Kromer embryo Klb (Normentafel, Xo. 3, 5 to 6 primitive somites). A., dorsal aorta; E., endothelial tube; M., myo-epicardial mantle; P., pericardial cavity; Ph., pharynx. X 100.

Fig. 373. — Model of the heart of a human embryo. No. 300 of Rob. Meyer's collection (Normentafel Xo. 7), 2.5 mm. greatest length. Modelled by P. Thompson. (After Thompson.) A., atrium; Au., region of the atrial canal; B., bulbus cordis; S., sinus venosus; T ., truncus arteriosus; U., vena umbilicalis sinistra; V., ventricular limb. X 50.

the heart cylinder becomes divided into its four portions. While the delimitation of the atrial portion from the ventricular limb is indicated by the constriction in the region of the atrial canal and that of the ventricular limb from the bulbus cordis by the constriction at the bottom of the bulbo-ventricufar cleft, that of the sinus from the atrium is less distinct. Later, however, this delimitation is made clearer by a groove which constricts the floor of the atrium from the left and delimits the left portion of the sinus venosus from the left portion of the transverse atrial sac. By this the wide connection between the sinus and the atrium is narrowed and the sinus itself is divided into a transverse middle portion and two lateral portions communicating with this — the transverse portion of the sinus and the left and right sinus horns.

538 The changes that now take place consist in a relative change of position of the individual portions of the loop. The sinus with its sinus horns, which up to this time has been the most caudal portion of the loop, comes to lie on the dorsal side of the transverse atrial sack ; at the same time the apex of the ventricular limb, which hitherto has looked towards the right, comes to be directed more caudally. This change of position of the bulboventricular limb becomes clear by a comparison of the position of the bulbo-ventricular cleft of Thompson's embryo (Fig. 373)

A . nsc.


V. om.

V. om.

Fia. 374. — Model of the heart of the embryo Halj of 3 mm. greatest length, 15 primitive somites. The property of the First Anatomical Institute, Vienna. Modelled by R. Weintraub. Seen from in front after removal of the anterior pericardial wall. The dotted lines represent the limits of the veins. A., atrium; A. asc, ascending aorta; Aran., amnion; Au., atrial canal; B., bulbus cordis; D., yolk-sack, cut at its margin; D. C, ductus Cuvieri; D. v., anterior communication of intestine with yolk-sack; M. a., anterior mesocardium; P., pericardium; S. 1., left sinus horn; V., ventricular limb; V. om., omphalomesenteric vein. X 150.

with that seen in Embryo Hal 2 (Fig. 374) ; in the former it is almost horizontal, in the latter almost vertical. By this change what was formerly the caudal limb of the loop becomes the left limb, and what was formerly the cranial one becomes the right. As this assumption of a vertical position by the ventricular limb progresses, the atrium rises so that it comes to lie in the dorsocranial side of the ventricular limb and a continually increasing portion of it becomes visible in an anterior view. Moreover, by this change the atrium comes into contact with the basal surface of



the bulbar limb, which bends backward almost in a horizontal plane, and as the atrium continues to grow forward it becomes slightly constricted by the bulbus. The movement of the sinus venosus in a craniodorsal direction, already described, accompanies this change in the atrium, so that the sinus, which formerly opened at the base of the atrial limb, now opens into the posterior wall of the atrium. At the same time, by the recession of the entire heart, the direction of the sinus horns is altered, these being directed no longer upward and medially, but assuming at first a horizontal direction and finally opening into the transverse portion of the sinus from above and laterally; they form

Ph. + Th

A. asc.





Fig. 375. — The model shown in Fig. 374 after removal of the pericardium, seen from behind. A., atrium; A. asc, ascending aorta; Amn., amnion; B., bulbus cordis; D. v., anterior communication of the intestine with the yolk-sack; E., ectoderm; Ph., lateral wall of pharynx; Ph. + Th., impression of the pharynx and the (median) thyreoid anlage on the pericardium; S. I., reflection of the pericardium upon the left sinus horn; S. r., right sinus horn; V., ventricle. Lateral to S. I. the transverse section of the left ductus Cuvieri is to be seen.

with the transverse portion an arch which is no longer convex upward, but is at first convex anteriorly and later downward. The bulbus cordis participates in the relative change of position of the various parts of the heart to the extent that its distal end becomes more and more pushed toward the median plane as the ventricular limb becomes more vertical, and at the same time its distal bend becomes straightened out. In earlier stages this bend is quite sharp, but later the slightly curved bulbus passes gradually into the truncus arteriosus.

This topical modification of the heart is accompanied with changes of form, which consist partly in the progressive delimi


tation of the four portions of the heart and partly in the further elaboration of each portion. Hand in hand with the change of position of the sinus there goes a diminution of the left sinus horn and a consequent retardation in the growth of the transverse portion of the sinus. The diminution of the horn is chiefly due to the obliteration of the left umbilical vein. The right half of the transverse part of the sinus is not greatly affected by the diminution, since in the meantime a number of hepatic veins have acquired openings into it. By the progressive constriction of the sinus from the posterior wall of the atrium from the left, the greatly narrowed opening of the sinus comes eventually to lie at the right posterior end of the transverse part of the atrium. The formerly transversely oval opening between the sinus and the atrium has been converted, by the developmental processes just described, into a longitudinally oval one, whose greatest diameter is placed sagittally and vertically. Toward the left it is bounded by the constriction described above, while on the right there is formed a vertical fold which, starting on the upper wall of the atrium, passes down the posterior wall beside the sinus opening to reach the lower wall. This fold is the first anlage of the right sinus valve (valvula venosa dextra).

The atrium changes its form to the extent that its two lateral extremities, enlarged in a balloon-like manner, project beyond the bulbar limb anteriorly and above, and it seems as if the right half decidedly surpasses the left in volume, at least the similar results of observations on several embryos of this stage point this way. The first division of the atrial cavity is due to the distal portion of the bulbus becoming lodged in its anterior upper wall. Corresponding with the blunt prominence projecting into the atrial cavity, produced in this manner, there develops, toward the end of the period now under consideration, a sickle-shaped fold, which extends at first over the posterior and later also over the anterior wall of the atrium, then gradually fading out. This is the anlage of the septum primum.

At about this same time the upper end of the right sinus valve thickens and projects into the atrial cavity as the first indication of the septum spurium; a slight groove on the outer surface of the atrium marks its position. Between this and the constriction produced by the lodgement of the distal end of the bulbus cordis, the cranial wall of the atrium is somewhat outpouched, forming the intersepto-valvular space. At the left end of the lower atrial wall is the entrance into the atrial canal, which at the beginning of this stage of development has a somewhat circular form, but at its end is transversely oval. At first this opening lies wholly to the left, associated with the abruptly descending left Avail of the atrium, but later it shifts

THE DEVELOPMENT OF THE HEART. 541 as a whole to the right, so that its right end i> finally situated at the middle of the atrial wall that is directed toward the ventricle. Corresponding to this change in the interior of the heart cavity a modification of the outer surface is naturally also visible. The atrial canal, which at fit st is visible on the left side of the heart, gradually comes to lie more and more deeply and vanishes towards the right, being overlapped from the left and above by the enlarging left atrium and from the left and below by the enlarging upper part of the ventricular limb. Hereby the atrial canal approaches nearer to the bulbo-ventricular cleft, until, finally, the depression bounding it on the left becomes continuous with the cleft and forms with it the bulbo-auricular groove. The ventricular limb, at this time, broadens in all directions and overlaps, especially on the left side, the circumference of the atrial canal. Its communication with the bulbus cordis, which at the beginning of this period was rather narrow, enlarges by the disappearance of the duplicature of the heart wall which is interposed between the ventricular and bulbar limbs, this duplicature being produced by the deep bulbo-ventricular cleft. In this process there is naturally no degeneration of heart substance, but merely a difference of growth to the disadvantage of the part under consideration. This lagging behind of the portion intervening between the bulbus and the ventricle shows itself on the outer surface of the heart by the bulbo-ventricular cleft becoming gradually shallower and gradually shortened in the caudocranial direction. The enlargement of the communication produces a common ventricular cavity, involving the transition portion, that is to say the caudal part, of the bulbo-ventricular limb. This becomes divided into two portions in its cranial part by the projection of the heart wall into the interior at the bottom of the bulbo-auricular cleft. This ridge-like projection, whose cranial portion was rather plump so long as the auricular canal lay entirely to the left, becomes sharper with the shifting of the atrial canal toward the right, and finally becomes a sharp-edged fold, which, as already stated, subdivides the cranial portion of the ventricular loop in the sagittal direction. To the left of this bulbo-auricular ridge lies the entrance to the atrium in the form of a well-defined atrial canal, to its right is the bulbus cordis, gradually diminishing in size as it is followed away from the ventricle. At the base of the common ventricular cavity there begins at this period the formation of a sagittally placed elevation, the first anlage of the interventricular septum.

The histogenetic processes which take place during the developmental period that has so far been followed are as follows. At first the distance between the endothelial cardiac tube and the myo-epicardial mantle is very great throughout the whole extent of

542 the heart anlage, and during life it seems to be filled with a serous fluid, since it is occupied in sections by a clot-like, fibrous mass, entirely destitute of cells and staining feebly with hematoxylin (Fig. 376). The endocardium consists of a layer of endothelial cells with large nuclei, while the myo-epicardial mantle is composed of several layers of cells, which have more of a syncytial character, at least their boundaries are distinguishable only rarely and sporadically. It is, however, difficult to determine how far this indistinct delimitation of the individual cells is due to the preservation or staining, since none of the embryos I have had for study were

--'•^**- lift-" 1


Fig. 376. — Transverse section through the embryo Hah. A. d., descending aorta; B., bulbus cordis; E., endothelium of the cardiac tube; H., auditory pit; M., myo-epicardial mantle; Ph., pharynx; V., ventricle. X 100.

stained with iron-hasmatoxylin. The space which at first exists between the endocardium and the myo-epicardium diminishes later in an irregular manner; it disappears first of all in the sinus and then in the atrium, so that in these regions the endocardium is in contact with the muscle mantle in early stages. In the region of the impaired ventricular cavity the apposition of the two layers occurs somewhat later, while throughout the circumference of the atrial canal and in the bulbus the apposition does not take place within the limits of the period now under consideration. In these regions there are formed in the space between the two

THE DEVELOPMENT OF THE HEART. 543 layers the so-called endocardial thickenings or endocardial cushions. In place of the absolutely cell-free, fibrous coagulum there occur in these regions sporadic stellate cells with relatively small nuclei, the staining with hematoxylin becomes decidedly stronger, and the whole tissue reminds one forcibly of the type of tissue seen in the Whartonian substance.

The myo-epicardial mantle differentiates to the extent that in the region of the ventricular loop and in that of the bulbus its superficial layer is formed by a continuous row of cells, the epicardium, while on the atrium and sinus, so far as the latter has a free surface, no such differentiation can as yet be said to occur. But the ventricular limb takes precedence over the atrium in the differentiation of the myocardium itself, as well as in that of the epicardium. In the ventricle one sees not only an increase E. Ma.



Fig. 377. — Section through the ventricular wall of the heart of embryo Hah, 3.5 mm. greatest length, in the collection of the I. Anatomical Institute, Vienna. E., endothelial cells; Ep., epicardium; M., myocardium; Mc, cortical and Ms., spongy portion. X 150.

of volume in the myocardium, but also its further differentiation and finally the appearance of trabecular These first appear at the base of the common ventricular cavity as projecting elevations, which gradually become more and more undermined, until finally, surrounded on all sides by the closely apposed endocardium, they traverse the ventricle more or less free. The elaboration of the trabecular network proceeds from the base toward the atrial canal on the one hand and toward the bulbus on the other, and at the close of this period one can speak of two portions in the cardiac musculature, an outer cortical and an inner trabecular or spongy portion (Fig. 397). The latter is of considerable thickness, but the corticalis forms only a thin layer and a difference in the degree of differentiation of the two portions is also perceptible. At the beginning of the period under consideration the entire myocardium stained deeply with eosin and the individual cells were rich in


protoplasm, but at this stage the spongy portion is composed of cells poor in protoplasm; at least these cells in well-preserved embryos stain feebly with eosin. On account of their poverty of protoplasm the boundaries of the individual myoblasts are more distinct than formerly. In the region of the trabecular myocardium there now appear in the otherwise feebly-stained cell-bodies fine, strongly eosinophile muscle fibrils, which do not confine themselves to individual cell territories, but traverse several cells. None of this fibrillar structure is yet to be seen in the cortical layer. The atrial myocardium, which, so far as its differentiation is concerned, behaves like the cortical layer of the ventricle, shows a discontinuity along the line of attachment of the posterior mesocardium, muscle substance being completely wanting in early stages along this narrow zone (area interposita of His). The right sinus valve in its early stages is a duplicative of the myocardium, at least it may be seen that embryonic connective tissue occurs between the two muscle lamellae. No boundary exists between the atrial and ventricular musculature, the former passing into the latter on all sides at the atrial canal. The myocardium is continued distally to the line of attachment of the pericardium, that is to say, to the region of transition from the bulbus to the truncus. Later this limit becomes less definite, as the distal portion of the bulbar myocardium apparently vanishes, a process by which the truncus arteriosus undergoes an elongation at the expense of the bulbus, and this at the time when the myoepicardial mantle is not yet differentiated at the distal end of the cardiac tube. This explains the difficulty which exists in determining the limits of these two portions of the efferent tube.

The endocardial thickenings in the atrial canal, mentioned above, may be described, in accordance with the form of the atrial canal itself, as an anterior endocardial cushion, situated on the anterior wall, and a posterior one on the posterior wall. On the small lateral borders of atrial canal the endocardium lies fairly close upon the myocardium. In the region of the bulbus a ring of endocardial tissue, at first of almost uniform thickness, replaces the space filled with fluid that is present in the earlier stages, and in this ring the following changes take place. In the proximal part of the bulb, the ventricular part, the endocardial thickening becomes especially strong along two spirally arranged regions, while in the intervals between these its development is retarded, and there are thus formed the proximal bulbar swellings, which, according to Born's method of nomenclature, may be termed the proximal bulbar swellings A and B. The distal portion of the swelling A lies on the left side of the bulbus, and as it descends it passes more and more toward the front, until, finally, at the proximal end of the bulb, it extends down toward the common ven



tricular cavity on the anterior wall. The proximal swelling B in its distal part lies on the right wall of the bulb and passes thence downward on the posterior wall to disappear in the posterior wall of the ventricular cavity, just as the swelling A does in the anterior wall. In the oldest embryos belonging to this period of development one sees already that the most proximal parts of both bulbar swellings are undermined by trabecular musculature ascending from the apex of the ventricle. In the distal or truncus portion of the bulb the endocardial ring is also in process of differentiation to the extent that in a series of sections one sees endocardial thickenings projecting toward the lumen to form the distal bulbar swellings. Distally the endocardial thickenings

A. d

Fig. 378. — Model of the heart of the embryo Hah of 5.2 mm. greatest length. In the collection of the I. Anatomical Institute, Vienna. Modelled by W. von Wieser. Seen from in front. A. d., right atrium; A. «., left atrium; An., region of the atrial canal; B., bulbus cordis; Bv., bulbo-ventricular cleft; T., truncus arteriosus; V., ventricle. X 100.

become gradually lower, until finally they pass over into the closely apposed endothelium of the truncus at the region where externally the indistinct boundary between the bulbus and truncus may be perceived. The projecting spur, the future septum aorto-pulmonale, which projects between the two halves of the system of aortic arches, i. e., between the • future systemic and pulmonary aortas, does not at this stage reach the line at which the pericardium is attracted to the truncus arteriosus.

In the second stage of development of the heart there is an approximation of the external form to the final condition, but the more important part of the progress is in connection with parts in the interior of the heart. As regards the external form, the Vol. II.— 35


546 change in the relative position of the various parts proceeds, the atrium gradually reaching a higher position, while the apex of the heart is carried so far caudally that, as is shown by a side view, it comes to lie caudal to the atrium. At the same time the opening of the sinus shifts completely to the dorsal wall of the atrium. The formerly slight constriction of the cranial wall of the atrium, produced by the bulb and the truncus arteriosus, now becomes a deep groove, and the lateral parts of the atrium on either side of this groove have enlarged so much that they begin to embrace the bulb as the anlagen of the auricular appendages (Fig. 378).

P.V.v.s. S.I V.p. p.




p.Bw. A


Fig. 379. — Model of the heart of embryo £U of 6.5 mm. greatest length. In the collection of the I. AnatomicalJInBtitute, Vienna (Normentafel No. 27). Modelled by J. Tandler. The lower half of the model divided transversely, seen from above. A. d., right atrium; A. «., left atrium; Ek. h., posterior endocardial cushion! of the atrial canal; P., pericardium; p.Bw. A, proximal bulbar swelling A; p.Bw.B, proximal bulbar [swelling B; S., sinus venosus; <S. /., septum primum; V., ventricle; V.p., vena pulmonalis (with sound inserted); V.v.d., right valvula venosa; V.v.s., left valvula venosa. X 100.

The posterior surface of the atrium also shows a shallow furrow, which corresponds to the oesophagus. To the right of the broad deep atrial furrow for the reception of the bulbus cordis, the groove corresponding to the attachment of the septum spurium, already described, has greatly deepened, so that a portion of the posterior wall of the right atrium projects in a dome-like manner, forming the spatium intersepto-valvulare. This is bounded below by a short transverse furrow, which separates it distinctly from the region in which the right sinus horn opens. The left sinus horn has emancipated itself from the pericardium to



the extent that at first it remains in connection with it only by a small band, resembling a mesentery, but in later stages this also vanishes and all connection between the sinus horn and the pericardium disappears. Similar conditions occur also in the transverse part of the sinus, except that in this region they occur somewhat later. The right sinus horn continues to enlarge without interruption and at the same time gradually ascends on the posterior atrial wall and is absorbed into the atrium, with the exception of its caudal portion, into which the transverse part of the sinus opens, and of its blind cranial end which is elevated in a



I Ek. h.


Fio. 380. — Model of [thejjheart of embryo La of 9 mm. greatest length. In the collection of the I. Anatomical Institute, Vienna (Normentafel No. 37). Modelled by J. Tandler. The atrial portion of the model has been divided frontally and the whole is viewed from in front. A.d., right atrium; A.s., left atrium; Ek.h., posterior endocardial cushion of the atrial canal; S.I, septum primum; S.I I, septum secundum; S.i., spatium interseptovalvulare ; S.iv., sulcus interventricularis; S.a., septum spurium; V.p. vena pulmonalis; V.v., right and left valvula venosa. X 75.

dome-like manner and is separated from the spatium interseptovalvulare by the groove already described (compare Fig. 380).

The changes taking place in the interior of the atrium may be described as follows (Fig. 379). The septum I, which grows downward from the posterior upper wall of the atrium and is at first quite low, becomes higher, and its ends, drawn out in a sickleshaped fashion, extend so far forward along the lower and upper walls of the atrium that they reach the margin of the atrial canal. At the same time the free edge of the septum becomes remarkably thickened and bounds, together with the plane of entrance of the

548 atrial canal, the primary narrowed opening of communication between the two atria, the foramen ovale I. Its line of origin, however, becomes gradually thinner and thinner, and finally there is formed, either directly at the line of origin of the septum on the posterior upper wall of the atrium or immediately below it, a dehiscence, the foramen ovale II, which rapidly enlarges. The septum I then gradually separates from its line of attachment and becomes a ribbon-like structure with fluted margins, traversing the atrium from behind and below forward and upward. In such hearts (compare embryo La, "Wal, Figs. 380, 381, 382) the original line of attachment of the septum I appears as a slight elevation on the dorsal wall of the atrium, and immediately to the right of




Fig. 381. — Sagittal section through the model shown in Fig. 380, the section passing to the left of the septum I. Seen from the left. Au., atrial canal; B., bulbus; F.o.II, foramen ovale II; S., sinus venosus ; S.I, septum primum; S.II, septum secundum; V., ventricle. Below the septum primum and above the atrial canal is the foramen ovale I.

this the elevation of the septum II begins; the further history of this may conveniently be described in the third period of development. In the first period the right sinus valve was the only one present, the left (valvula venosa sinistra) being scarcely indicated, but now the latter is strongly developed. The two valves lie one on either side of the slit-like opening of the sinus, which is directed from above and outward, inward and downward (Fig. 379). The left one unites on the cranial wall of the atrium with the thickening* of the right valve, which has already been described, and forms with it a large distinct septum-like structure which passes over the cranial wall of the atrium on to the anterior wall and is



the septum spurium of His (Fig. 380). Caudally the two valves behave differently, in that the right one gradually flattens ont on the floor of the atrium, while the left one extendi toward the sicklelike end of the septum II, which is growing backward on the floor of the atrium, and later unites with it. 3 The space between the left valvula venosa and the anlage of the septum II is outpouched dorsocranially to form the spatium intersepto-valvulare. To the left of the septum I, in the angle between the posterior and lower

.•V-':".\ ; .".'.:'.'.


Fig. 382. — Transverse section through the heart region of the embryo Wal of 8 mm. greatest length. In the collection of the I. Anatomical Institute, Vienna. A.d., descending aorta; Au., atrial canal;!"S., sinus; S.I, septum primum; V.v., valvule venos».

walls of the atrium and in the region where externally the posterior mesocardium is still attached, is the opening of a. vessel coming from the lungs, the single vena pulmonalis (Figs. 379, 380). In a section through the model of a heart at this stage (Fig. 380) one sees how distinctly the atrium has become separated from the ventricle by the deepening of the atrioventricular groove, — that is to say, how greatly the atrium and ventricle have become ex

3 The details of this process will be described with the next period of development.


panded beyond the outline of the atrial canal. Changes in the position and form of this canal have also taken place. As regards its position it is to be" noted that it has made such progress in its shifting toward the right that it has already come to lie in the centre of the floor of the atrium. If in the model of a heart at this stage one looks from the ventricle into the atrium through the atrial canal, one sees that the septum I is directed exactly toward the centre of the transverse diameter of the canal (Figs. 379, 380). As the result of this further shifting toward the right the bulbo-atrial ridge, already described, becomes still more prominent and simultaneously with the shifting other changes take place in the canal. In correspondence with the bulging of the ventricle beyond the circumference of the canal, which has already been noted, and, further, in correspondence with the continued undermining of the endocardial cushions by the musculature, the endothelial swellings project freely some distance further into the ventricular cavity. Both the anterior and the posterior swellings become modified in such a way that their lateral extremities become more elevated, while their central parts remain somewhat flatter; consequently one may distinguish in each endocardial cushion a middle straight portion and two lateral elevations or tubercles. The shape of the atrial canal m transverse section thus comes to resemble the figure formed by two T's placed base to base ( I— H). This peculiar modification of the atrial canal is completed relatively quickly; small endocardial thickenings also appear on the lateral margins of the canal in later stages.

In the ventricular limb on the convexity of the common ventricular cavity, — that is to say, at the point where in the earlier stage the ventricular limb passed into the bulbus, — a constriction appears, extending over the ventricular surface of the heart and gradually becoming shallower as it is traced upward. This interventricular groove (Fig. 380) marks externally the separation between the right and left ventricle and divides the blunt apex of the heart into two portions, so that at this stage the right and left ventricle each has its own apex. The portion belonging to the left ventricle is, however, greater than that pertaining to the right one. Corresponding to this external groove, the interventricular septum, already seen in the earlier period as a rounded ridge, becomes more prominent, but just as there is externally an asymmetry in the two ventricles, so too in the interior the subdivisions of the ventricular cavity are by no means equal at first. This inequality of the ventricles is later partly compensated for by a broadening of the right one, but a slight asymmetry persists in that the interventricular septum is so placed that its prolongation would not cut the middle of the atrial canal but the right tubercules of the endothelial cushions.

THE DEVELOPMENT OF THE HEART. 551 The most proximal portion of the bulbus has enlarged considerably and has been taken up into the ventricle, and at the same time the greater part of the bulbo-ventricular cleft has disappeared. Processes having an important bearing on the entire subdivision of the heart take place in the bulb during this period of development. Attention has been called to the fact that already in the first period endocardial thickenings develop in the bulbus ; these were termed the proximal bulbar swellings A and B, and, as their name indicates, they are situated in the proximal half of the bulbus. They have, as has also been stated, a spiral course around the inner surface of the bulbus, yet they have grown more distally, so that the swelling A, beginning distally on the left posterior wall, passes thence to the left, to finally disappear proximally on the right anterior bulbar wall, as this ascends from the d.Bw.2 mil mf

d. Bw. 1

ft .'*" VV.vS^-v-Tf

d. Bw. 3


Fia. 383. — Section through the bulbus cordis of the embryo H6. The section cute the bulbus obliquely to its long axis, and consequently the line of attachment of the pericardium is also cut. d.Bw. 1-4, distal bulbar swellings 1-4.

common ventricular chamber without any sharp delimitation from it. The swelling B begins distally on the anterior wall of the bulbus and passes thence over the right wall to the posterior one, where it disappears at about the same level as swelling A, at the junction of the bulbus and ventricle. The conditions at the proximal ends of the bulbar swellings will be described in detail later on. In earlier stages an endocardial thickening occurred around the whole circumference of- the bulb in its distal portion, and a differentiation of the distal bulbar swellings had not yet taken place. In the period now under consideration these are developed ; but it may be said that they do not present the regularity of form and occurrence that obtains in the birds and reptiles. To demonstrate the extent of the proximal and distal bulbar swellings it is convenient to divide the bulbus into a proximal

552 ventricular and a distal truncus portion; these subdivisions can only be temporary, however, since the bulbus during this period of development undergoes a continuous and rather rapid shortening at both ends. Its central end is gradually taken up into the right ventricle, while the truncus arteriosus elongates heart-ward at the expense of its distal end. In this shortening bulbus the proximal bulbar swellings occupy the proximal half and the distal swellings the distal half ; yet this delimitation is not quite accurate, at least for the distal bulbar swellings 1 and 3, to be described below, since these gradually pass over into the proximal swellings A and B. Four distal bulbar swellings can be distinguished, and starting with the right distal one and proceeding to the left and backward they may be denoted by the numbers 1-4. When fol

S.'a. p.

d. Bw. 3




p. Bw. A

Fig. 384. — Model of the bulbus cordis of the embryo H6, divided longitudinally. The left half of the model is shown. -4., aorta (4th aortic arch); d.Bw. 1-3, distal bulbar swellings 1-3; P., attachment of pericardium; p. Bw. A, B, proximal bulbar swellings A, B; PL, pulmonary artery (6th aortic aroh); S.a.p., septum aorto-pulmonale; *, point at which the sound in the lumen of the pulmonary artery disappears, being covered by the fusion of the distal bulbar swellings 1 and 3, forming the distal bulbar septum; **, point at which the sound again appears in the common lumen. Proximally the aorta and pulmonary artery are separated by the two proximal bulbar swellings coming into contact to form the proximal bulbar septum. The subdivision of the common efferent tube is produced distally by the septum aorto-pulmonale, in the middle region by the distal bulbar septum and proximally by the proximal bulbar septum. Between these three portions of the partition there are two points of communication, in which the ends of the sound are visible.

lowed proximally they are seen to run downward on the bulbus walls in a clock-wise spiral. They do not project equally into the lumen of the bulbus, but swellings 1 and 3 are strongly developed while 2 and 4 are weaker. Swelling 1 lies distally on the right wall of the bulb and passes gradually backward and to the left, swelling 3 begins above on the left wall and passes to the right anterior one as it descends; thus it is possible for them to pass over into the proximal swellings A and B in later stages, since swelling A passes distally on to the left posterior and swelling B on to the right anterior wall of the bulbus. The swellings 2 and 4 have a position between swellings 1 and 3, swelling 2 passing from above and behind downward and to the left and swelling 4 from above and in front downward and to the right.

THE DEVELOPMENT OF THE HEART. 553 In addition to these two sets of bulbar swellings the septum aorto-pulmonale, in so far as it is a derivative of the truncus wall, must be considered in connection with the subdivision of the efferent tube. The partition between the sixth and fourth pairs of aortic arches, which in earlier stages reached to the line of attachment of the pericardium, grows proximally in this period of development and extends into the portion of the efferent tube that is already intrapericardial. At the same time a continually increasing portion of that part of the truncus which was originally outside the pericardium is brought within its territory by the elongation of its walls at the expense of those of the bulbus. The processes by which this change of the walls is brought about will be described later. Three factors, accordingly, take part in the subdivision of the efferent tube, the septum aorto-pulmonale and the distal and proximal bulbar swellings. At the beginning of the second period of development — in embryro H G (Fig. 384), for example — these three portions are still distinctly separated. The septum aorto-pulmonale ends bluntly, and with its prolongations are associated the distal bulbar swellings 1 and 3, which for a certain distance still project but little into the lumen, so that in this region the aorta and pulmonary artery still remain in communication ; more proximally the two swellings are in contact and consequently separate the two arterial tubes. They terminate immediately below this region of contact and are still distinctly separated from the proximal swellings A and B. At this point the more or less broad, single lumen of the proximal half of the bulb begins, and the proximal swellings, which project extensively into the lumen of the bulb, gradually flatten out. In the succeeding stages of this period the septum aorto-pulmonale reaches the point of union of the distal bulbar swellings 1 and 3, so that the aorta and pulmonary artery become separate throughout the entire extent of the distal half of the bulb; yet even in this stage the limit between the septum aorto-pulmonale and the distal bulbar septum, as the union of the distal bulbar swellings 1 and 3 may be termed, may be recognized by the differences in the histological structure of the two partitions and of the walls of the truncus and bulbus.

As regards the tissue differentiation in this period two distinct processes may be recognized : first, the differentiation of the mvocardium. and, second, the continued development of the endocardial thickenings. The differentiation of the muscular tissue proceeds more rapidly in the ventricle than in the atrium. In the latter the tissue occurs in the septum T and also in the two sinus valves, the myocardium of these latter being a single structure and no longer appearing as a duplicating projecting into the atrial cavity. It is continued as a strong bundle into the septum spurium, and.

554 in addition, there are present some muscular ridges projecting into the lumen of the atrium, the first anlagen of the musculi pectinati. The right sinus horn, the transverse portion of the sinus, and even the left horn possess a musculature, and in the walls of the atrial canal the atrial musculature at all points is continuous with that of the ventricle. In this two portions may again be recognized, a peripheral cortical and a central trabecular layer, the latter being everywhere more differentiated than the former. In sections in which the trabecular musculature is cut longitudinally (Fig. 385) it may be seen that the fibrillar have become quite long and occupy the entire breadth of the prismatic cells ; correspondingly the boundaries between individual cells are

^> *

Fig. 385. — Section through the wall of the ventricle of embryo EU (Normentafel No. 27). Mc, cortical substance; Ms., spongy substance. Muscle fibrilte are distinctly visible in the spongy substance. X 150.

still quite distinct where they are in contact by their lateral surfaces, while in those places where their bases are in contact the boundaries have vanished, at least it is impossible to say, on account of the extensive development of the fibrillar, to what extent these belong to one cell or the other. In transverse sections through trabecular the cell boundaries are therefore plainly visible, but at the same time the greater part of the cell body is already occupied by fibrillae. In sections of the cortical portion one can see that only the peripheral portions of the cells are occupied by very fine fibrils, the central portions being free from them and poor in protoplasm. While the development of the cortical substance in the proximal part of the bulbus keeps pace with that

THE DEVELOPMENT OF THE HEART. 555 of the ventricle, the musculature of the distal part is less differentiated. The myocardium of this portion surrounds the bulbus tube as a distinct muscle layer and extends peripherally as far as the distal bulbar swellings can be traced, the layer, however, gradually becoming thinner distally and the differentiation of the myo-epicardial mantle less pronounced, until finally it appears not only as if a further differentiation of it had not occurred, but even as if degeneration had taken place. All those parts of the efferent tube in whose lumen the partition is formed by the septum aorto-pulmonale are destitute of myocardium.

The cardiac endothelium rests smoothly on the subjacent tissue in all parts of the heart and, as in earlier stages, consists of a single layer of flat cells with relatively large nuclei. The endocardial thickenings have the following distribution: in the atrium the free, thickened edge of the septum I is provided throughout its entire length with a small endocardial thickening, whose anterior and posterior prolongations extend as far as the corresponding endothelial cushions of the atrial canal and fuse with them. Consequently the foramen ovale primum is completely surrounded by an endocardial thickening. The shape of the endothelial cushions of the atrial canal has already been described, but it may be remarked that the masses of tissue now stain more deeply with hematoxylin and their nuclei are more abundant. The endocardial swellings on the narrow sides of the atrial canal have also been described already. The most important change that occurs in the atrial endothelial cushions is their undermining by the trabecular musculature of the ventricle. In earlier stages- the slope of the endothelial swellings toward the ventricular cavity was a gradual one and they did not overhang the ventricle; but now, partly by the downgrowth of the endocardial cushions and partly by the undermining of their attachment, they project freely into the lumen of the ventricle with sharp edges. The undermining is brought about by the continued extension of the trabecular network, which may be followed as far as the attachment of the anterior and posterior endocardial swellings. In the atrial canal itself the cortical and spongy substances are not yet differentiated and as a single sheet pass over into the atrial musculature. The lateral endocardial thickenings have not yet been affected by the undermining process.

The proximal bulbar swellings are bounded externally in their upper parts by the muscle ring of cortical substance ; if, however, they are traced proximally they show a change at about the level of the atrial canal, in that there develops between the endocardial thickenings and the cortical substance of the bulb a system of trabecular musculature, which is at first thin and composed of scattered trabecular, but increases gradually in thickness toward


the ventricle. Here also the endocardial swellings become undermined by trabecular, and in places where the bulbar swellings have ceased for some time to be distinguishable in the model as elevations directed toward the ventricle, one sees as final prolongations of them, endocardial thickenings on the surface of the trabecular network which looks toward the lumen of the ventricle. These conditions show, moreover, the extent to which the progressive absorption of the bulbus into the ventricle has advanced in given cases. Histologically the bulbar swellings differ from the endocardial cushions principally by being somewhat poorer in cells. The distal bulbar swellings are similar in structure to the proximal ones, but the septum aorto-pulmonale is quite different. Here one finds a connective tissue which is very rich in cells with large nuclei and which does not differ from that of the rest of the wall of the aorta and pulmonary artery. This tissue also does not stain diffusely with hematoxylin. Where the septum aorto-pulmonale ends — that is to say, where it passes over into the distal bulbar swellings 1 and 3 — the histological character of the wall alters, a very delicate ring of but slightly differentiated myocardium making its appearance.

At the conclusion of the period of development just described the subdivision of the heart into the right and left halves has advanced so far that the anlagen of almost all portions of the cardiac septum have appeared and the individual cardiac cavities communicate only by more or less wide openings. In the succeeding third period of development the subdivision is completed, with the exception of that of the atria, which, as is well known, only becomes perfect post partum. But this period, which includes embryos from about 10 to 20 mm. vertex-breech measurement, and extends from the fifth to the eighth week of fetal life, shows not only the completion of the subdivision of the heart but also the almost complete development of the valve apparatus. At the close of the period the outer form of the heart and the subdivision of the ventricular cavity and bulbus are complete, but only in the next and last period is the final development of the interior of the atrium accomplished and the histological differentiation of the heart then reaches its completion. This period extends to the close of fetal life and, indeed, is not quite completed at birth.

Beginning with the changes that take place in the sinus during this period, it is seen that simultaneously with the gradual retrogression of the left sinus horn, the right one sinks more and more to the level of the posterior wall of the atrium, until finally it no longer is seen rising above the posterior surface of the atrium when the right atrium is viewed from behind. This, however, is not due to a fusion of the atrial and sinus walls, but to the absorption of the sinus walls into the posterior atrial wall by the flat

THE DEVELOPMENT OF THE HEART. 557 tening out of the furrows bounding the sinus and by the passive stretching of the sinus wall, which lags behind the rapidly growing atrium, so that both the transverse and vertical diameters of the sinus are enlarged. Thereby the opening of the superior vena cava (ductus Cuvieri), which has appeared in the meantime, is shifted from the posterior to the upper atrial wall, and similarly the inferior cava is shifted to the inferior wall. The portion of the sinus wall situated between these two vessels becomes at the same time part of the posterior wall of the atrium, and this exogenous portion of the atrial wall is delimited from the parts in its neighborhood by the line of attachment of the two sinus valves. While the right sinus valve is still very high at the close of this period, the left one lags behind in its development and undergoes a modification to be described later.

With this absorption of the sinus wall into the posterior wall of the atrium there occurs a change in the opening of the transverse portion of the sinus into the right sinus horn. This transverse portion, the continuation of the left horn, opened hitherto into the left lower angle of the right sinus horn immediately beside the opening of the inferior vena cava. By the absorption of the sinus the opening is brought to the level of the posterior atrial wall, and with it the spur-like elevation between it and the right hom, representing the former bend of the transverse portion toward the right horn. This sinus septum, that seems to arise from the posterior wall of the atrium, now grows so far toward the right that it reaches the valvula venosa dextra and divides this into two portions, — a shorter portion in front of and below the line of meeting of the two structures, and a longer portion behind and above, passing upward over the posterior wall of the atrium and disappearing in the septum spurium. When the last period of development is being considered it will be seen that from the former portion the valvula Thebesii is formed and from the lower part of the latter portion the valvula Eustachii.

The foramen ovale I closes at the end of the preceding period by the fusion of the free edge of the septum I with the endocardial cushions of the atrial canal, but the foramen ovale II still forms a wide communication between the two atria on account of the feeble height of the septum II. Later this septum increases in height and there is in consequence a narrowing of the foramen ovale II. But in addition another change occurs in the circumference of the foramen, dependent upon a change in the direction of growth of the two septa. This change is as follows: at the beginning of this period the free edge of the septum I, directed toward the foramen ovale II, in the natural position of the heart (the plane of the foramina atrio-ventricnlaria almost frontal), looks backward and upward ; gradually, however, its lower pro


longation extends backward and upward, at first over the posterior wall of the atrium and finally over the upper wall, while the upper prolongation lags behind in its growth, so that now, with the heart in the same position, the free edge of the septum looks forward and upward. At the beginning of this period the septum II is still low, and its free edge, directed toward the foramen ovale II, looks forward and downward in the natural position of the heart. Later the septum becomes higher and at the same time its anterior prolongation grows forward and downward over the upper wall of the atrium, until finally the free edge of the septum looks backward and downward. The two septa have thus altered their relative positions to the extent that the posterior prolongation of the septum I on the left side has grown past the line of attachment of the septum II, and the anterior prolongation of the septum II has similarly on the right side grown past the line of attachment of the septum I. The left sinus valve now also takes part in the formation of the circumference of the foramen ovale II in the following manner : the outpouching of the right atrium, the spatium inter septo-valvulare, described in the preceding period of development and situated between the left sinus valve or the septum spurium and the septum atriorum, continually lags behind in its growth. Consequently the prominence produced by the spatium on the posterior wall of the atrium also disappears and the valvula venosa sinistra gradually approaches the septum atriorum, and, finally, there remains of the once extensive spatium intersepto-valvulare only a small cleft-like recess, which is closed below by the fusion of the lower end of the left sinus valve with the lower prolongation of the septum secundum. Later, while the destruction of the spatium intersepto-valvulare is taking place by the fusion of the valvula venosa sinistra with the septum I, the upper and middle portions of the left sinus valve vanish more or less completely, but the lower part, persisting on account of its union with the septum II as described above, elongates its free border backward and upward, and so completes later the limbus Vieussenii, which is formed from this free border.

In the earlier period of development the single quite short pulmonary vein trunk opens close to the line of attachment of the septum atriorum. Later there is an absorption of this short trunk into the posterior wall of the atrium, so that the right and left pulmonary veins open into the left atrium by two separate openings. The portion of the atrial wall between the two openings has therefore been formed by an originally extracardial portion of the pulmonary veins, and later it increases rather rapidly in breadth, so that the two pulmonary veins become separated more and more.

The changes in the form of the atrial canal have been followed

THE DEVELOPMENT OF THE HEART. 559 to the time when the canal is a slit so narrow in the frontal direction that its central portion is a mere cleft, while its lateral extremities represent the places in which the atrio-ventricular valves will develop. At the beginning of the present period of development the opposed edges of the slit fuse throughout their whole extent, where the marginal tubercles, that have already been described, occur. In this way the single atrial canal is divided to form the two atrio-ventricular orifices, which are separated by the entire width of the zone of fusion. The septum primum rests upon this zone of fusion.

The subdivision of the ventricular cavity has progressed by the ventricular septum becoming higher, so that only a small opening, the remains of the foramen interventricular e, exists between its upwardly concave margin and the under surfaces of the endocardial cushions, which, in the meantime, have fused. The anterior end of the ventricular septum, which if prolonged would come into relation with the two right tubercles of the endothelial swelling, passes without interruption into the remains of the bulbo-atrial ridge, while the posterior prolongation applies itself directly to the right tubercle of the posterior endocardial cushion.

Before the processes which lead to the final closure of the foramen interventriculare are described it will be necessary to consider in detail the subdivision of the bulb, since the two sets of processes not only take place simultaneously but also show a causal dependence. After the fusion of the distal bulbar swellings 1 and 3 the lumen of the aorta contains one half of each bulbar swelling 1 and 3 and the whole of the swelling 4, while the pulmonary artery has the other halves of swellings 1 and 3 and the entire swelling 2. The external groove between the aorta and pulmonary artery, which was present in earlier stages only in the distal portion, has now become prolonged proximally and has deepened, so that the two vessels have almost circular lumina.

While the peripheral portions of the distal bulbar swellings flatten out and finally disappear, their most proximal portions, which have come into relation with the proximal swellings A and B, not only retain their former height but increase in size and begin to be hollowed out in their distal slopes. Thus there are formed in each vessel three plump folds directed distally, the first anlagen of the semilunar valves. The pouch-like cavities between these folds and the walls of the vessels gradually increase in size, partly by the folds becoming thinner and partly by the outpouching of the corresponding portions of the walls of the vessels, and the evaginations so formed are the anlagen of the sinus Valsalva. A complete separation of the aorta and pulmonary artery has thus been accomplished, and the semilunar valves develop, as has just been described, from the lower end<= of the


distal bulbar swellings. Proximally the proximal bulbar swellings A and B, which have increased in height in the mean time, fuse to form a short septum, the proximal septum bulbi, which extends toward the ventricle from the line of the semilunar valves. This septum lies in the same plane as the septum between the aorta and pulmonary artery, which was formed by the fusion of the two distal swellings 1 and 3, but it becomes replaced by the tissue of the septum aorto-pulmonale as the wall of the truncus arteriosus elongates proximally at the expense of the wall of the bulbus. The proximal septum thins out rapidly as it is traced downward, and, in the natural position of the heart, it extends from the right above

FL r.H.v.E.

p.S.B. / E.v. IH.v.E.

// / i

I. II. h. E


Fig. 386. — Model of the heart of embryo S2 of 14.5 mm. greatest length (Normentafel No. 58). In the collection of the I. Anatomical Institute, Vienna. Modelled by J. Tandler. The model has been divided transversely midway between the atrio-ventricular groove and the apex of the ventricle and the upper half is shown from below. E.r., the two endocardial cushions fused; l.H.h.E., left tubercle of the posterior cushion; l.H.v.E., left tubercle of the anterior cushion; O.v.d., right ostium venosum; O.v.s., left ostium venosum; PL, pulmonary artery; p.S.B., proximal septum bulbi; r.H.t.e., right tubercle of the anterior endocardial cushion; S.v., septum ventricularum; Y.b.l.Z., lateral cusp of bicuspid valve. The arrow points to the orifice of the aorta. X 44.

and behind to the left down and forward, and consequently does not lie in line with the sagittally placed septum interventriculare, but forms with it a sharp angle, open upward and backward (Fig. 386). The proximal septum bulbi has a free border that is concave downward, and its anterior prolongation, the bulbar swelling A, becomes continuous with the anterior prolongation of the upwardly concave septum interventriculare, while its posterior prolongation, the swelling B, deviates to the right of the right tubercle of the anterior endocardial cushion and becomes greatly broadened and flattened; the posterior prolongation of the septum interventriculare, however, runs toward the right tubercle of the posterior endo

THE DEVELOPMENT OF THE HEART. 561 cardial cushion. If one follows the margin of the foramen hiterventriculare, beginning with the posterior prolongation of the septum bulbi, it is found to be a spiral ridge, that runs first upward along the free concave border of the proximal septum bulbi, passing anteriorly into the edge of the septum interventriculare, then along this downward and backward and finally upward again, to terminate at the right tubercle of the posterior endocardial cushion. The opening so bounded unites not only the two ventricles, but also leads upward and to the right into the pulmonary artery and to the left and upward into the aorta. It becomes closed in the following manner: the proximal septum bulbi grows downward and reaches the septum interventriculare. The right ends of the endocardial cushions, which have fused in the meantime, undoubtedly participate in the fusion of the two septa, but the extent of their participation cannot be exactly determined, since the entire circumference of the foramen interventriculare is surrounded by endocardial growths which pass into one another. Since from the very beginning the septum interventriculare tends toward the right tubercles of the endocardial cushions, and since the final closure of the foramen interventriculare takes place in the region where they occur, it follows that the right atrio-ventricular orifice lies immediately adjacent to the point of closure, while the left one is separated from it by the entire breadth of the fused endocardial cushions. A further complication is introduced by the fact that the closure of the foramen takes place with the aid of the prolonged septum aorto-pulmonale, so that the right portion of the circumference of the aorta comes to lie at the very point of closure. Consequently the septum membranaceum, which is formed at the point of closure, forms a constituent of the wall of the conus arteriosus aortce.

Since the septum atriorum is attached to the endocardial cushions much further to the left than the septum interventriclare. the portion of the fused cushions between the attachments of the two septa, when it later comes to lie in the planes of the septa, does not separate ventricle from ventricle, but the sinus arteriosus aortae, i. e., the left ventricle, from the right auricle, and it may therefore be termed the septum atrio-ventriculare (Hochstetter). The distance between the semilunar valves and the point of final closure of the foramen interventriculare is still relatively great, so that the aorta, and especially the pulmonary artery, arise from the ventricles as elongated cones. The inner surfaces of the cones are smooth, trabecular musculature not having yet developed in their walls. "With the closure of the foramen interventriculare the final subdivision of the ventricular portion of the heart is completed.

The histological changes occurring during this period may be Vol. II.— 36


described as follows : The left sinus horn possesses far distally a muscle mantle that is interrupted by the openings of certain cardiac veins. The differentiation of this musculature, as well as of that in the region of the atrial wall which has been formed by the absorption of the right sinus horn, lags far behind that of the rest of the atrium. In the regions of the septum I and the septum II, as well as in the sinus valves, musculature is present, but it has made no especial progress in differentiation. In the portions of the atrium which represent the auricular appendages trabecular have developed, more apparently on the right side than on the left. In the small area of the left atrium, which was formed by the absorption of the pulmonary trunk, no musculature is evident v. v.




S. a. v..



Fig. 387. — Section through the heart of embryo Mi of 16.75 mm. greatest length. In the collection of the I. Anatomical Institute, Vienna. Semidiagrammatic. S.a., septum atriorum; S.a.v., septum atrioventriculare; S.v., septum ventriculorum; V.d., right ventricle; V.s., left ventricle; V.v., valvule venosse.

at this period. As to the ventricular musculature it may be noted in the first place that the wall of the left ventricle undoubtedly surpasses that of the right in thickness. The difference depends, however, especially on the cortical substance. The trabecule are exceedingly numerous and almost fill the entire cavities, yet they are not so uniform in thickness and differentiation as in the earlier stages, but trabecular conspicuous by their strength and advanced differentiation occur in both ventricles. These trabecule, from their topography and from their relation to the valves, are to be identified as anlagen of the musculi papillares. While, at the beginning of the period under consideration, the cortical substance was relatively thin, at the end of the period it has become greatly developed and its fibrillar structure has made further progress.

THE DEVELOPMENT OF THE HEART. 563 In longitudinal and transverse sections through the cardiac musculature the individual muscle-cells show distinct cell boundaries on their long sides, but pass into one another on their short sides without any delimitation, a condition that was present to some extent in the earlier period of development, but has now become general. To this extent it is proper to speak of a cardiac syncytium at the end of this period. 4 At the same time the myocardium shows a further differentiation in that the individual rows of muscle-cells have already become arranged to form muscle bands, just as one sees them in the adult heart.

While up to the present the atrial and ventricular musculatures were continuous around the atrial canal, one now sees these two portions of the musculature lose their continuity. In earlier stages there was along the line of the atrio-ventricular groove an aggregation of embryonic connective tissue, which was wedgeshaped in section, the edge of the wedge being directed inward. In later stages this wedge gradually becomes prolonged toward the cavity of the heart and cuts so deeply in between the atrial and ventricular musculature throughout the entire circumference of the atrio-ventricular groove that the direct continuity of the atrial and ventricular cortical substance is interrupted. Thus the original continuous cortical musculature is divided into two portions, a sinus-atrial portion and a ventricular portion, while the trabecular parts are still in continuity at the anlagen of the valves.

The changes in the form of the endocardial thickenings occurring in the atrial canal and in the bulbus have already been described. The following points may be noted as regards their texture. The plump semilunar valves, formed from the proximal ends of the distal bulbar swellings, do not yet show any special differentiation of their tissue, at least their diffuse staining has not appreciably diminished, although the nuclei of these portions of the endocardial swellings are perhaps somewhat more closely set than formerly. The part of the septum aorto-pulmonale that follows, i e., the proximal septum bulbi, still shows all the characteristics of endocardial growths, as do also its prolongations, the remains of the bulbar swellings A and B. If one follows the bulbar swelling A, one sees the tissue characteristic of such swellings fade out at the free edge of the septum interventriculare and pass posteriorly without interruption into the fused endothelial swelling. The tissue of the posterior bulbar swelling becomes 4 The old discussion concerning the boundaries of the cardiac muscle-cells cannot be considered here. As has already been described, the fibrillas pass beyond the cell territories, a condition which may justify the term syncytium. The condition described above is not, however, sufficient for the settlement of the question, since the staining of the objects (haematoxylin-eosin) is not suitable for final conclusions.


greatly broadened and in part passes medially into the tissue of the endocardial cushions and into the endocardial thickening on the free edge of the interventricular septum, and in part it passes laterally toward the lateral circumference of the right atrio-ventricular orifice and may be followed for some distance toward the apex in the posterior wall of the ventricle.

When the closure of the interventricular foramen is completed, nothing can be determined as to the origin of the various parts contributing to the closure from their texture, the entire region being occupied by an endocardial tissue, relatively poor in cells and staining diffusely with hsematoxylin, and which at the very spot of the future septum membranaceum undergoes rather soon a further development ; it loses its diffuse staining with haematoxylin and at the same time becomes richer in cells. The anlagen of the valve cusps, which arise from the anterior and posterior endocardial cushions, as well as from the endocardial thickenings at the lateral ends of the atrial canal, lag behind the region just described in their differentiation (Fig. 387). These anlagen are plump and their undermined borders are connected with the trabecule of the ventricles. This connection is quite distinct where especially strong Irabeculae, the anlagen of the musculi papillares, come into connection with definite portions of the valves. In such places it is possible to follow muscle bundles ascending from the cortical substance through the entire length of the papillary anlage to the valve. The muscle bundles on the atrial surface of the valve anlagen which have been described as passing toward the ventricles are apparently continuous with the ventricular bundles just described; at least no boundaries between the two could be made out. Thus the valve cusps are for a time a series of plump elevations, consisting partly of endocardial growths and partly of musculature; a differentiation into actual musculi papillares, chordas tendinese, and valve flaps cannot be distinguished at this stage.

In addition, the degenerated processes which occur in the region of the bulbus cordis are of interest. At the beginning of the period under consideration the undermining of the proximal bulbar swellings by the trabecular myocardium has not extended verv far. The older the embryo, the more the trabecular tissue grows toward the line of attachment of the semilunar valves. On the other hand, one sees the cortical substance, which formerly surrounded the bulbus far distally, gradually receding, the recession being accompanied by a simultaneous change in the structure of walls of the derivatives of the distal part of the bulbus, the aorta and pulmonary artery. The typical wall of the two arteries extends gradually proximally, until finally this typically layered wall, destitute of myocardium, mav be followed to near the semi

THE DEVELOPMENT OF THE HEART. 565 lunar valves. The part of the bulb immediately distal to- the valve zone thus becomes surrounded by a common myocardial layer, so that in this stage of development the future bulbus aortae is surrounded for a considerable distance by cardiac musculature. Toward the end of the period this portion of the musculature is so far degenerated that the relation of the aorta to the musculature is almost that which obtains in the heart of the child.

In the following period of development the parts of the heart that have been already elaborated undergo a further modelling and are slightly altered, but one sees nowhere any extensive transformations of the constituent parts of the heart. The same is true also with regard to the further histological differentiation.

As regards, first of all, the sinus portion of the heart, one sees that, in consequence of the apposition of the left sinus valve to the septum atriorum, the latter comes to represent the medial boundary of the sinus area, while the lateral boundary is formed by the derivatives of the right sinus valve or of the septum spurium. It is therefore necessary to consider first the extensive modifications of the right sinus valve (Fig. 388). As was noted in the account of the preceding period of development, this broad valve, which projects markedly into the lumen of the atrium, is divided by the apposition of the sinus septum into two parts, an anterior inferior (ventral) and a posterior superior (dorsal). The dorsal portion of the valve is continued without interruption into the septum spurium, which formerly served as a common stay for the two valvular venosae, and later it undergoes further modifications. The portion of valve corresponding to the posterior wall of the atrium, which was also originally high, gradually flattens and forms a prolongation of the ridge-like elevation of the septum spurium on the upper wall of the atrium, carrying it over this wall backward and downward. This small ridge, which forms the persistent lateral boundary of the original sinus area, consists in its upper portion of the rudiment of the septum spurium and in its posterior inferior portion of the rudiment of the right sinus valve, the crista terminalis of His.

The lower part of the dorsal portion of the sinus valve remains high and bounds the opening of the inferior vena cava laterally, as the valvula vence cavce Evstachii. Frequently one may actually speak of an increase in the surface of this portion, but the valve presents great variation in its height, development, and form. In accordance with its development the valvula Eustachii is continuous above and behind with the crista terminalis and in front and below it ends at the border of the transverse portion of the sinus, i. e., the sinus coronarius cordis.

The ventral, much smaller portion of the right sinus valve bounds the opening of the sinus coronarius, that has been formed

566 from the transverse portion of the sinus, and eventually becomes the valvula Thebesii.

Since the atrial septum must furnish until birth a means of communication between the two atria, a final closure of this portion of the heart cannot occur during embryonic life. In the developmental period under consideration the septum I increases in height more rapidly than its surrounding parts increase in extent ; the communication between the two atria thus becomes somewhat narrowed, but, on account of the oblique position of the septum I, it continues to be wide enough to allow free passage of the blood from the right atrium into the left. The oblique position of the septum I is not determined by any special mode of growth of the


S. c. S. s.

Fig. 388. — Model of the heart of an embryo of 310 mm. greatest length. Modelled by Born. C. i., vena cava inferior, in which a sound is placed; C. t., crista terminalis; S.I, septum primum; S. II, septum secundum; S. c, opening of coronary sinus; S. s., sinus septum; Y. E., valvula Eustachii; V.Th., valvula Thebesii.

septum, but is produced mechanically by the greater pressure of the blood in the right atrium. The septum I, i. e., the valvula f oraminis ovalis Vetteri, only attains its definitive position when the blood pressure becomes equal in the two atria, and then only does it have an opportunity for fusing with the septum II and so bringing about the final separation of the two atria.

Before mentioning the changes that take place in the ventricles and the valve apparatus, a few words are necessary as to the modifications that now occur in the left atrium. By the absorption of the originally single pulmonary vein trunk, these veins are now represented by two stems, and at the same time a new portion,

THE DEVELOPMENT OF THE HEART. 567 correspondiDg to the distance between the two pulmonary orifices, has been added to the atrial wall. This portion continues to increase in breadth, and at the same time the absorption of the two pulmonary vein stems progresses until these have been taken up into the atrial wall as far as their first division. As a result the originally single openings of the right and left venae pulmonales become again divided, so that on either side there is an upper and a lower pulmonary vein orifice. The portion of the atrial wall between each pair of pulmonary veins was likewise originally a part of the wall of the veins. It will thus be seen that the participation of the pulmonary veins in the formation of the wall of the left atrium is quite extensive.

The anatomical changes occurring in the ventricular portion of the heart during this period can be briefly described. The valve apparatus alone undergoes further elaboration in that the differentiation of the valve cusps, the chorda tendinece, and the musculi papillares makes further progress and the formal differentiation of the semilunar valves increases.

The histological differentiation of the heart approaches its conclusion during this period. The principal changes occur in connection with the valve apparatus, but before this is described it will be necessary to consider briefly the remaining portions of the heart. And, first of all, it may be noted that even in the latest fetal stages the development of the sinus musculature lags behind that of the rest of the atrium.

While the part of the right sinus valve which becomes the valvula Eustachii continues to lose its musculature, that of the crista terminalis continually increases. The rudiment of the left sinus valve still shows traces of musculature in later stages of fetal life, as, for instance, in a fetus of 150 mm. vertex-breech length. The originally thin margin of the limbus Vieussenii thickens by the development of strong muscle bundles in it.

The portion of the left atrial wall lying between the openings of the pulmonary veins was destitute of musculature in the earlier period of development, but in this period it acquires a complete muscle layer. The differentiation of the musculi pectinati takes place more rapidly than that of the remaining portions of the atria.

The greater thickness of the cortical substance of the left ventricle continually increases (Fig. 389), and the fibrillar of this layer become more and more abundant until the difference between the cortical and spongy substances which obtained at the beginning of the period now under consideration gradually disappears. In this period also loose subepicardial connective tissue appears on the surface of the heart along the lines of the vessels and nerves.

The semilunar valves become thinner, their connective tissue


loses its succulent character, becoming fibrous and tendinous, and finally presents the characteristic appearance seen in the child.

While in the earlier period the endocardial portion of the atrio-ventricular valves greatly surpassed the muscular, at the beginning of the present period the trabecular which pass from the spongy substance of the ventricle toward the valves increase greatly, and at the same time the central portions of the valves come to project further into the lumen, owing to the growth of their peripheral portions. In accordance with this process the trabecular of the spongy substance, which originally were almost immediately in contact with the cortical substance in the neghborhood of the atrio-ventricular orifices, separate from the corticalis more and more and pass centrally. In this stage also not one free-ending trabecula could be found as an indication of a secondary union between the valves and the papillary muscles. At this time also

V. d.

Fig. 389. — Section through the heart of an embryo of 165 mm. greatest length. V. d., right ventricle; V. s., left ventricle. The difference in the thickness of the two ventricles is apparent.

the musculature on the atrial surfaces of the valves becomes greatly developed, so that, in comparison with later stages, it actually seems as if the valves were exclusively muscular in structure, with the exception of the purely endocardial marginal portions which are directed toward the lumen. Later the musculature degenerates, at first in the valve areas, and connective tissue takes its place, standing in intimate connection with the connective-tissue wedge which was described as occurring in the earlier stage of development and which separates the cortical substance of the atria from that of the ventricles. Still later one sees that the peripheral portion of the ventricular valve musculature degenerates, at first at the surface and then more deeply, giving place to connective tissue and so becoming transformed into chordce tendinece. In later stages accordingly one must distinguish between a connectivetissue valve (the secondary valve of Bernays), chordce tendinece, and papillary muscles. The free margins of the valves at certain points retain for a longer time the characteristic structure of the endocardial thickenings, while in other points they are converted into typical connective tissue. These thickenings represent the noduli Albini.


The developmental history of the atrio-ventricular system is at present practically unknown. The time allowed for the completion of the present article did not permit a thorough study of the question and a satisfactory solution of it, and I shall therefore state the results of my observations briefly.

In a human embryo of 19.75 mm. one sees at the upper border of the septum inusculare, close to the lower surface of the not yet completely differentiated endothelial swelling, a triangular area, which occupies the tip of the muscular septum and is distinguishable even under a low magnification by its special staining properties. Its nuclei are dark and the cell bodies stain faintly with eosin. It is not unlike a sympathetic ganglion, but its cells are less numerous. This structure corresponds in position with the stem of the His bundle, and it already possesses a right and left prolongation, which are the right and left limbs of the atrio-ventricular system.

An embryo of 28.5 mm. shows a similar condition at the same region, but the two limbs have become longer and the cells larger. A similar aggregation of cells is to be seen in the region of the septum atriorum, immediately above the septum membranaceum.

In the study of the development of the bundle of His the following point is of importance.

The conducting system may either be a persistent connection between the atrial and ventricular musculatures situated in the posterior wall of the heart, in which case it would represent an ancient connection between the two parts of the heart, or it may be a new development which has been formed only after the completion of the septum, in other words only secondarily. If the latter be the case, the conducting apparatus of hearts without a septum is quite a different affair from that of hearts which possess a septum. Furthermore the conduction of stimuli in hearts which possess a septum must be a different affair before the development of the septum — that is to say, before the development of the bundle of His — from what it is later on.

So far as my observations go, I incline to the view that the His bundle does not represent the persistence of an ancient atrioventricular connection. This view has also been expressed by Retzer, who has supplied some data as to the development of the atrio-ventricular system of the pig.


Bernats, A. C. : Entwicklungsgeschichte der Atrioventrikularklappen. Morphol.

Jahrbuch. Vol. 2. 1876. Born, G. : Ueber die Bildung der Klappen, Ostien und Seheidewande in Sauge tierherzen. Anat. Anz. Vol. 3. 1888. Beitrage zur Entwicklungsgeschichte des Saugetierherzens. Arch, fiir mikr.

Anat. Vol. 33. 1889.


His, W. : Anatoniie mensehlieher Embryoneu. Leipzig. 1880-1885.

Mitteilungen zur Embryologie der Saugetiere und des Menschen. Arch.

fur Anat. u. Phys. 1881. Ueber die Entwieklung der Form und der Abteilungen des Herzens. Comptes Rendus Congres period, intemat. des sci. med. 1884. I. Section d' Anatomie.

Copenhagen. 1886. Beitrage zur Anatomie des menschlichen Herzens. Leipzig. 1886. Hochstetter, F. : Ueber die pars membranaeea septi. Vortrag gehalten in der wissenschaftlichen Aerztegesellsehaft zu Innsbruck. Wiener klin. Wochen schr. 1898. Die Entwieklung des Blutgefasssystems. Hertwig's Handb. der vergl. und experim. Entwicklungsgesch. d. Wirbeltiere. Vol. 3. 1906. (Published 1901 and 1903.) Keibel, F., and Elze, C. : Normentafel zur Entwicklungsgeschiehte des Menschen.

Jena. 1908. Mollier: Die erste Anlage des Herzens bei den Wirbeltieren. Hertwig's Handb.

d. vergl. u. exper. Entwicklungsgesch. d. Wirbeltiere. Vol. I. 1906. Retzer: Some Results of Recent Investigations on the Mammalian Heart. Anat.

Record. Vol. 2. 1908. Spee, Graf F. : Beobachtungen an einer menschlichen Keimscheibe mit offener Medularrinne und eanalis neurentericus. Arch, fiir Anat. u. Physiol. Anat.

Abth. 1889. Thompson, P. : Description of a Human Embryo of 23 Paired Somites. Journ.

of Anat. and Physiol. Vol. 41. 1907.


By HERBERT M. EVANS. Johns Hopkins University, Baltimore.

1. General.

We shall consider here, first, the more general questions concerning the development of the vascular system, and, secondly, the special development of the vascular system in human embryos.

In recent years studies on the vascular system of the higher vertebrates have opened up new and profitable fields and have given us a better conception of the method by which the blood-vessels grow and become transformed in the general growth of the embryo.

In the following account I shall confine myself to the history of the chief vascular trunks only, 1 for here our knowledge now stands on a firm footing, and any laws which we may discover as applicable in these instances may safely be taken as of general worth.

The two fundamental questions involved in the development of the vascular system are — 1. What is the origin of the blood-vessels in the body of the embryo? 2. What is the primitive form of the vessels in any area, and the manner of change from this to that of the adult? These two aspects of the subject thus concern themselves with the problem of the cellular antecedents of the endothelium, on the one hand, and with the principles governing the architecture of the vascular system, on the other.

To the former problem it is still impossible to give any decisive answer, but to the latter I trust the reader will see that a flood of new light has come.

1 There exist few accounts of the development of peripheral vessels, but mention mav be made of the work of Mall, Flint. Miller. Sabin, and Fnchs.

DEVELOPMENT OF THE VASCULAR SYSTEM. 571 1. Human embryos, as will be mentioned further on, have contributed little information on the origin of the cells forming the vascular system, and indeed after a wealth of observations on other animals this question is still a very open one, having met with a decisive answer in no case.

In embryos of the higher vertebrates the first cells which can be identified as standing in any relation to the vascular system are in the form of localized thickenings of the extra-embryonic mesoderm 2 lying next the endoderm of the yolk-sac. These constitute the so-called vascular anlagen, and typically undergo a gradual differentiation from a nest of indifferent cells into two more definite cell types, blood-cells on the one hand and endothelium on the other. The endothelial cells enclose the former, and, continuing to divide, produce vascular sprouts and thus extend themselves into new areas. While this differentiation of the earlier anlagen is progressing, new anlagen are formed by the mesoblast, but eventually a time is reached when this latter process ceases, and subsequently in the history of the embryo all endothelium is derived from that of pre-existing vessels. That this is the case in older embryos and in the adult has been verified by many observations. It is important, then, to distinguish vessels which have arisen through the sprouting of the endothelium of other vessels in contrast to vessels whose endothelium has been contributed directly from the neighboring mesoderm. Even on the yolk-sac these latter vessels which arise in loco are not numerous, for they only occur at the site of the so-called anlagen, and the main mass of the vitelline capillary plexus arises from the extension and frequent anastomoses of these primary vessels.

It is a question now whether the early blood-vessels in the body of the embryo itself are not formed by an ingrowth of the vitelline capillaries, or whether, on the other hand, the embryonic stems, or at least a part of them, do not arise in situ from the mesoderm of the body. Both of these positions have been defended, the name of His (1900) being identified with the former idea and that of Riiekert and Mollier (1906) especially with the latter.

In the birds it has been possible to establish beyond all doubt that most of the aorta descendens is formed from the medial margin of the vitelline capillary plexus (Vialleton 1892, His 1900, Evans 1909; see Eig. 390). The frequent early connections of this vessel with the same plexus in mammals makes it highly probable that a similar origin obtains here (Tursting, 1884). For the head portion of the aortae, on the other hand, conflicting accounts are given. His described it as arising from a continued growth of the same extra-embryonic capillaries which formed the vessel in its lower course, but which were restricted to a capillary chain growing headward, eventually turning ventrally over the blind end of the head-gut and fusing with the cephalic portion of the heart tube. On the contrary, Riiekert and Mollier have given various details of what they consider the local origin of the aorta in this locality from the mesodermal cells of the lateral plate of the mesoderm and from the splanchnic mesoderm. It is not possible then to state positively that the yolk-sac anlagen are the only source of the endothelium of the body vessels, for the earliest of these latter may themselves 1 Although lying in the mesoderm, these anlagen may have actually arisen from the entoderm. This is the view taken by Ruekert (1906), who has been the last to subject the question to a special study. On the other hand, most investigators, beginning with Kolliker's early work on the rabbit (1875), have emphatically denied any entodermal participation here, and affirmed that the blood islands of mammalian embryos are to be looked upon as special localized proliferations of the mesoblast. Robinson (mouse) and Heape (mole), Janosik (marmot, pig), Fleischman (cat), Keibel (guinea-pig, man), and Van der Strieht (rabbit and bat) may be mentioned.

572 be primary vascular anlagen in the sense of being directly derived from neighboring mesoderm. Another possible source for the endothelium of the vessels of the head is constituted by the paired anlage of the heart (C. Rabl, 1887). 2a After the establishment of the aorta it is possible to satisfactorily deny the further local origin of any of the subsequent vessels of the embryo, since these can all be demonstrated to arise from capillary sprouts of true vascular


f \

K *<

Fig. 390a.

Fig. 3906.

Fig. 390a. — Ventral view of the left side of a rabbit embryo of five somites, showing the vascular plexus from which the left heart and aortae are derived. The heart is already indicated as an especially enlarged member of the mesh. This is not yet true for the aorta; the arrows indicate two places where the aorta is as yet unconnected into a longitudinal plexus. The brackets show the position of the five somites. ( After J. L. Bremer.) Fig. 3906. — Ventral view of posterior portion of an injected chick embryo of 20 somites, showing the formation of the lower aortae from a capillary plexus continuous with that of the yolk-sac.

endothelium, just, as is the case in every locality where the development of vessels has been carefully studied in the living animal, e.g., the tail of the living frog larva.

The various claims for a local origin of blood-vessels relatively late in the growth of the embryo have gradually been successfully disproved. It will be remembered that the appearances known to many observers as Ranvier's " vaso formative cells " were supposedly instances in which a local origin of blood-vessels occurred relatively late in the growth of the embiyo. Recently Vosmear (1898) has shown

" a Since the above was written, Bremer (1911) has demonstrated that in the head of rabbit ernbryos of five somites, the aorta is represented by a distinct plexiform angioblast coterminous with the vitelline angioblast. These facts make it highly probable that the extra embryonic endothelium has grown into the body in this region just as can be demonstrated in successive stages for the more caudal portion of the aorta of the chick.

DEVELOPMENT OF THE VASCULAR SYSTEM. 573 that the vessels in question were isolated secondarily after having clearly arisen from other vessels, and Clark has observed the same phenomenon in the living frog larva (personal communication). (See also Renant, 1901, 1902.) It would seem that a mere histological analysis, even though on perfectly fixed material, would not suffice to settle the question of the delicate connection of embryonic vessels. These collapse so readily that the most perfect of the usual methods of study will not suffice to disclose them. On the other hand, injections of the vascular systems in young embryos show a wealth of capillaries and interconnections not hitherto demonstrable, and it would consequently seem unwise to overvalue any negative evidence in this respect given by uninjected embryos.

It is evident, then, that, while it is probable that the only source for the endothelium for the blood-vessels is comprised in the cells of the vascular anlagen, it is nevertheless possible to prove that this source is comprised in the endothelium of the first intra-embryonic vessels (aortae and cardinal veins), however these may have arisen. Injections of the embryo after these early stages and subsequent exploration with the microscope show no vessels unconnected with the general system, and lead us to be certain that new vessels in any area arise exclusively as offshoots from the older ones. This doctrine of the specificity of the endothelium has met many apparent confirmations in the histogenesis of new growths, for there also, as in normal development, Rabl's dictum is doubtless true, " Endothelium only from endothelium." 2. Concerning the development of the form of the vascular system, two positions have been taken, — one, that the arteries and veins grow out as single trunks to their respective territories, the other, that the first vessels in any area are capillaries usually in the form of a typical plexus from which secondarily arteries and veins arise.

It will be seen that a correct conception of the actual truth here affects vitally our ideas even of the factors concerned in development as a whole ; for it is difficult to see in the blind outgrowth of single trunks to their future territory anything but a teleological design or hereditary predestination. On the other hand, the adherents of the idea of a capillary plexus ancestry for vessels view the vascular system as functioning from the beginning, and the formation of arteries and veins as only an expression of the functional adaptation of these plexuses to a beating heart.

The immense number of vascular variations in the adult, which seem to take every conceivable direction, and the occurrence of arterial and venous anastomoses, early led the Swiss anatomist Aeby (1868) to suppose that the vascular system, arteries as well as veins, existed originally in the form of a uniform mesh-work of vessels, in which, so to speak, a competition took place for supremacy, and, the victors being the only trunks remaining, we obtained the dendritic branched appearance of the adult vascular system. Occasionally the primitive net was retained, and in these instances we saw retia mirabilia, or merely anastomoses between vessels.

In the case of variations the theory was most convenient, for the hypothetical uniform net furnished the possibility for a vessel to course in practically any direction. At a loss for a better explanation, Krause (1876) adopted the Aeby hypothetical plexus to account for the vast range of blood-vessel variations which he recorded in his well known chapter in Henle's Handbuch. But until recently Aeby's ideas have met with no other favorable reception. Indeed they were strongly opposed by the careful work inaugurated by the rise of a more exact comparative anatomy, in which C. Gegenbaur and his pupils are to be mentioned. Extensive comparative investigations soon showed that there was to be observed everywhere a remarkable constancy in the number and position of the vascular trunks and their relation to other structures (muscles, nerves).

Ruge's (1883) epoch-making work in this field demonstrated clearly that when variations occurred they tended to group themselves into quite definite types, which could be explained by the over-development of. normally inconspicuous

574 vessels, " collateral stems " or aberrants. Thus the old conception of the outgrowth of single trunks was only strengthened, for there could be considered present in anomalies only an unusually, strong outgrowth of a normally small trunk. Moreover, another authority in this field, F. Hochstetter, took occasion to denounce the Aeby-Krause idea. In his study of the developing vessels in the early limbs, Hochstetter declared he could find no instances of an indifferent condition of the vascular system anywhere in the limb buds, and that consequently " die Hypothese Baader's und Krause's als vollkommen unrichtig bezeiehnet werden muss." s (Hochstetter, 1891, p. 42.)


-Umbilical Vein


--Pnm. Cap. Plexus

Prim. cHJbCl.



Vein --19*D.I.V

Fig. 391. — Injection showing the profuse outgrowth of primary subclavian capillaries into the early wing bud of a_chick 60 hours old. 14th D. I. V., fourteenth dorsal intersegmental vein.

In 1894 R. Thoma published the results of a study he had been conducting on the ancestry of the vascular trunks present in the chick's yolk-sac. Thoma set himself the task of solving how it came about that arteries and veins were 3 This statement was at least partially justified by the simple conditions Hochstetter saw in the vessels of the limbs and tail of Triton, for here the vascular trunks are remarkably simple and suffer a more or less direct transformation into arteries and veins. That such simple conditions do not apply to the limbs of higher vertebrates (birds and mammals) the studies of Goppert (1910) and myself (1909) will demonstrate.



differentiated in this locality. Early stages showed him only an indifferent network of vessels in which no predominate trunks could be distinguished, and he was able to secure successive preparations showing the gradual formation of arteries and veins from this capillary net. The elaboration of these larger supplying and draining vessels represented merely a functional adaptation of the net to the demands of the circulation and a fortuitous location with regard to the aorta? or the venous ostia of the heart determined the use and enlargement of certain channels of the net to become arteries and veins respectively.

Thoma's ideas did not at first attract the notice they deserved. Nevertheless, Mall (1908) indicated that they were applicable in the development of the

Aorta dorsalis dext.

card. post, dext.

27th dorsal segmental vessel

Fig. 392. — Injection into the early leg buds of a chick of 32 somites, showing the capillary plexus.

body's vascular system no less than in the extra-embryonic area vasculosa, and in succeeding years he and his pupils repeatedly furnished evidence that such was indeed the case. Flint's (1903) study of the submaxillary gland and more recently of the lung (1906) showed that a capillary net always existed at the periphery of the growing vascular tree, and Mall (1906) instanced the same phenomena in the division and growth of the lobules of the liver. It may be noted that Zuckerkandl (1894) had reported an exactly similar phenomenon in the development of the median artery of the arm, for the early stages in the history of this vessel were represented by a chain of capillaries accompanying the median nerve.


In 1903 E. Miiller reported the results of a study of the development of the vessels in the human arm. His reconstructions showed that in some instances what is undoubtedly a true arterial plexus may exist in the region of the axillary artery.* These appearances were quite impossible to harmouize with the old idea of the outgrowth of single vascular trunks and led Miiller consequently to dispute this prevailing notion.

Four years later H. Rabl published the results of his study of the early vessels in the wing bud of the duck, and showed clearly that many "of these could be detected in arising from parts of a capillary plexus. Rabl also established the origin of the secondary subclavian artery, which characterizes the birds, from a chain of capillaries which grows caudally from the third aortic arch to join the plexus of the wing bud.

Above all, now, the admirable study of the vessels in the developing arm of the white mouse which E. Goppert has recently published shows clearly the development cf the successive branches of the subclavian artery "auf der Grundlage eines capillaren Netzes," and I cannot doubt but that in the further growth of the vascular system, in the various regions of the body, we will be able to observe these facts again and again.

Very recently methods of injecting living embryos so that the delicate vascular system is completely filled and yet extravasations avoided, have yielded a wealth of facts on the development of the vessels. The revelations due to such preparations have enabled us to see the capillary precxirsers of some of the more fundamental vascular trunks of the body.

Thus, if injections of chick embryos are made just preceding and during the time in which the limb buds are beginning to be elevated from the body wall, it is possible to trace the earliest vascularization of the limbs. Such preparations show that a series of capillaries springs from the lateral aortic wall opposite the limb eminence, and, anastomosing together, form a typical capillary plexus in the early limb tissue. The preservation and enlargement of one of these many aortic offshoots constitute the subclavian and sciatic artery respectively (Figs. 391 and 392).

Still other large vessels in the body can be traced to a similar stage in which there exists only a simple capillary plexus, out of which the main vessel arises through the utilization of a single channel in the mesh and the coincident atrophy of the remainder. Thus, in the early stages the head of the embryo possesses as its only vessels a delicate plexus of capillaries which arise at many points from the aorta? and more caudally are connected with the vitelline vein. Eventually with the circulation through this primary head capillary plexus, arteries and veins are formed from some channels in the mesh, and it is exactly in this way that the main stems of the internal carotid artery and jugular veins are formed (Figs. 393, 394, 395).

The pulmonary arteries are represented at first by a capillary plexus which arises from the sixth aortic arches and grows caudallv on to the lung bud (Fie. 396).

In the chick it is easy to see that the gut arteries are earliest represented by a plexus of capillaries which arise from the ventral aortic wall.

With all these instances, however, of an early plexiform anlage for many 4 A fact which has more recently been confirmed for the mouse by the important researches of Goppert (1909) in the same territory. The exact significance which Miiller would attach to this axillary plexus must be disputed, as also his contention for its constant occurrence. It must be considered merely one of the instances where the circulation has for a time taken equally favored paths through a preceding capillary plexus and thus formed for us for a time several anastomosing embryonic arteries rather than a single one, which is normally the case.



vessels, we are forced to admit that some of the primary vessels of the body are not preceded by such stages, but from the very first occupy a definite position and consist of only a single endothelial tube. The most striking example of this

Primary head plexus

V. cardinalis ant."

Ear pit.- ^>

1st segmental vessel

Eye vesicle

Later ductus Cuvieri V. vitellina

Fig. 393. — Lateral view of head of an injected chick of 15 somites, showing the primary capillary plexus here. The plexus takes origin from the convexity of the first aortic arch, and is continued posteriorly as a slender capillary chain which eventually joins the main vitelline vein near the junction of the latter with the heart. This slender capillary chain has arisen at several points from the dorsal aorta on each side, and two of these points of origin are still preserved opposite the region of the hind-brain. The delicate capillary path from head to vitelline vein is destined to form the anterior cardinal vein.

is furnished by the dorsal segmental arteries, which, as is well known, arise from the aorta at strictly intersegmental points and are usually distinctly single vessels.

The sharply limited definite positions of such great vessels as the aortae and umbilical veins are also phenomena which have been known for a long time Vol. II.— 37

578 and which seem unquestionably due to inheritance. However, even in these cases, an exacter study shows that these vessels do not develop at first as merely simple tubes. When, for instance, we turn to a consideration of the aorta, we can see clearly that in the chick the lower part of this vessel is merely the exaggerated medial margin of the vitelline capillary plexus, which has invaded the embryonic tissue (Fig. 390). This is also exactly the condition which may be seen in human embryos of corresponding age, for here also the caudal part of the aorta is

Capillary plexus of the head

Capillaries from the internal carotid artery

1st aortic arch

Anlage of the ventral part of the 3d aortic arch

Fig. 394. — Lateral view of injected pig embryo of 5.7 mm. length, showing the capillary origin of the a. car. int. X 80.

only a part of a general plexus of vessels which lie in the gut wall and continue to grow caudally. The dorsalmost members of this plexus straighten out longitudinally and form the aorta dorsalis while the connections of the latter with the plexus become the primitive vitello-umbilical complex of arteries. Inasmuch as in embryos of 6 somites these conditions occur opposite the future 7th and 8th somite (i. e., cervical somite 5), it is hence apparent that all of the aorta which exists later below this level has emerged from this preceding stage. There is little doubt but that the cephalic portion of the aorta has also an identical development, although less study has been given the appropriate stages. Turstig (1884) long ago showed that this was at first not a single vessel but a narrow meshwork of vessels



(Fig. 397), and that the single aortic tube came about only secondarily by the enlargement of one channel of the narrow mesh or by the fusion of several capillaries in some places. In the latter instances he described very clearly remnants of the old partition walls between the individual preceding channels in the form of cross-strands joining the ventral and dorsal aortic wall. Recently, Bremer (1911) has demonstrated the plexiform endothelial anlage of the entire cephalic portion of the aortas in rabbit embryos of five somites. Lingering remains of the several capillary channels which constitute the aorta in its earliest stages are occasionally

Ventral capillary network (Anlage of a. basilaris)

V. cardinal, ant.

- 1st aortic arch Ear vesicle

2d aortic arch

Capillary branches to the peduncle of the optic vesicle Internal caroid artery

3d aortic arch

Fig. 395. — Lateral view of injected pig embryo measuring 6.5 mm. The injection was made while the heart was still beating and shows the extent of the primary capillary plexus in the head. X SO.

seen in somewhat older specimens, where it is not uncommon to find the aorta splitting into two or three vessels to reunite again, a fact which I can confirm as occurring now and then in the human embryo with from 6 to 8 somites (Embryos Pfannenstiel-Kroemer, Etemod, Graf Spee).

The aortic arches are similarly formed by narrow chains of capillaries which quickly give way to the employment of a single channel in the mesh, though in some instances remains of the earlier plexiform condition persist. I figure here the first aortic arch in a young duck embryo (Fig. 39S).

It is only natural that the studies which have previously been made on the vascular system have usually revealed only the chief stems and have consequently led us to suppose that these stems grew out as such, for the usual methods of reconstruction of uninjected embryos can not hope to reveal more than the

580 chief trunks in any area. I present, for example (Figs. 399, 400), an embryo of about the same degree of development as Elze's (1907) shown in Fig. 420. His is from an unusually good reconstruction, mine from an injected specimen. It may be well to note that in the area here figured the reconstructed figure gives the appearance of the anterior cardinal vein growing forward and dorsally to constitute the future superior sagittal sinus; but the injected embryo shows clearly that this structure is beginning to be formed by the enlargement of the medial dorsal margin of the capillary mesh here, somewhat as the lower aorta? represent the enlarged medial margin of the vitelline net.

Most important of all, then, is the fact that the injections show its that the vascular system is not merely growing in an irregular fashion to obey an impulse given by heredity, but that it constitutes a connected and functioning whole.

Truncus arteriosus

Right pulmonary artery

A part of the auricular wall

6th aortic arch

Left pulmonary artery

§r— Blind branch to the left lung

Venous orifice Fia. 396. — Pulmonary vessels in a guinea-pig embryo 21 days old. (After Fedorow, 1911.)

If now we assemble the remarks which have just been made it is evident that we may state generally that arteries and veins do not grow out as such, but that the blood-vessels tend always to be laid down in a multiple capillary anlage rather than in single trunk-like forms, and that this is true even where tUe position of the vessel is apparently predetermined by inheritance. In many areas, however (e. g., the head and the limbs), we have more typical plexuses from which, through the secondary enlargement of some channels in the mesh and the coincident atrophy of others, arterial and venous vessels develop.

These facts are in accord with what we know to be the manner of development of blood-vessels in areas which are open to direct observation in the living animal. I refer, for example, to the studies which have been made ever since the time of Schwann on the tail of larval amphibia, where the transparency of the tissue



enables one to see the various structures in their growth and to prove for himself that new vessels are formed by endothelial buds and that the latter in turn form plexuses. It is only necessary to refer here to the classical observations oi' Kolliker (1S46), Remak (1850), Billroth (1856), Strieker (1865), Golubew (1S69), Arnold (1S71), Rouget (1873), Bobritzky (1SS5), Clark (1909), and others. Clark, in a piece of careful work on the tadpole, has now proved that we can thus actually watch the outgrowth and transformation of the primary plexus in an area so large that we may note that what are at one time parts of the most peripheral members of the capillary plexus are actually later used to become the arterial pathwa} r s for capillaries which have extended far more peripheralward.

Fig. 397. — Graphic reconstructions of the blood-vessels present in three stages of early rabbit embryos, showing the formation of the aorta. (.After Tiirstig, Schriften herausgegeben von der NaturforschenGesellschaft an der Universitiit Dorpat, I, 18S4.) A, B, and C, embryos possessing 3, 4, and 7 somites respectively.

I may also point out that this method of blood-vessel formation and growth has also been demonstrated in all cases where it has been carefully studied in the adult, — e. g\, in the vascularization of granulation tissue, new growths, etc.

The cause of the early appearance of vessels in a multiple capillary form is consequently to be found in the view that this represents the fundamental method of vascular growth, and that larger vessels only come into existence secondarily when the number of capillaries induces an increased supply of blood. Such an event leads to the enlargement of certain fortuitously situated capillaries into arteries and veins. The larger vessels are to be considered in the light of servants of the capillaries, for which they are but the delivering and draining pipes. Consequently the cause for the rich vascularity of a tissue cannot be sought in its possession of larger vessels, but rather in the influences which have brought about a more abundant growth of capillaries in it.

582 It may be noted now that, in addition to the method of capillary sprouts and plexuses, the blood-vessels in some special regions may be looked upon as arising in an essentially different way. I refer to the invasion of large venous trunks by certain tissues in such a way that the trunk becomes broken up into a great number of smaller vessels which now nourish the tissue in question. The fundamental point here is that we have capillaries interposed in a strong venous stream instead of between arteries and veins. The best examples of this are furnished by the invasion of the vitelline veins by the liver tissue, which thus breaks these vessels up into portal and hepatic sj 7 steins, and the invasion of the posterior cardinal veins by the mesonephric tubules, creating a transient renal-portal system (F. T. Lewis, 1904). The vessels formed in this way are markedly irregular and often

Branch encircling the optic vesicle

1st aortic arch in the plexus stage

V. vitellina dextra ,


Fig. 39S. — Injection of a duck embryo possessing 13 somites, made while the heart was still beating.

Viewed ventrally.

much larger than normal capillaries. So striking in fact is the picture produced by vessels which have arisen in this way and so many are the points of difference with the usual capillary plexuses that Minot (1900) has designated them sinusoids. It will be necessary now to refer briefly to the capillary plexuses occurring in the development of the embryo and the relation of these to the tissues. It may be stated first of all that no one has been able to verify the exact conception of Aeby, according to which a homogeneous mesh of vessels pervades all the tissues of the body. This is, mdeed, almost as far from the truth as is the existence merely of isolated arterial and venous channels. All recent work has shown that definite vascular and non-vascidar areas exist in the embryo, and that the capillaries grow from a vascular area into an adjoining non-vascular one. Thus, in the beginning the entire embryonic body is non -vascular, and after the formation of the aorta? we can recognize vascular centres or areas from which the capillaries continue to spread into areas which are as yet non-vascular. But the capillaries do not spread evenly in their growth from centre to periphery, thus invading




p Be

a CD

tl •d a a>


— o B < to <-»> a p •-»

3 o s CO £ a

CD o ire o


S' •i co Ci

»-i o a 00 p s " B in JC*

1 3.



I p O c —

HJ /.

CO s.


5' c P

GO B rv p

bb B



-1 S p S.

o B «J_ <

5' CD O 03 e* —

— >* o a " —1 B 3 B" -.


3' a ee 03 p B' o P .

o p O Wl B P p B &l ~ -• o < a


5 c 03 CD 0.

a •g ] 5' X ifs' CD a •"1 •1 B a* O a* <-t

  • i

b? B «< 5' Z

o » a «<; a B — ' m B P Bi K H

5' CD


B ^ CD CD P « B O 5' P CD •1 w B ef a.

B cr

p o o

o rt


a' p c < O B.

a ^*

P w .

a P


P -'

H n


CD c*


  • .

— O B ^ a < o


03 CD h" — C


p P <J

  • J

d P o

»l p

CD •-1

P a.

o 5'


CD Ert

►1 03



— a B CD

o 2,

•1 5'



'- \



quite uniformly an ever-widening zone, but, on the contrary, are apparently governed, even from the beginning, by the nature of the tissues, some attracting them early and others relatively late. Thus, the central . nervous system is early supplied by a close capillary net ; other areas in the embryonic tissue are apparently inimical to capillary growth, and these constitute distinctly limited non-vascular nones. Of these are to be mentioned those early condensations of the mesenchyme which represent pre-muscle and pre-cartilage masses. We are not improbably dealing here with the question of a chemical stimulant or " tropism " for endothelial proliferation, and may consider some tissues as possessing marked angiotactic properties in contract with a corresponding lack of them in others. It will be recalled that some tissues — e. g., the articular cartilages and cornea — remain nonvascular in the adult.

Ail the branches of an artery do not necessarily arise from the same primitive capillary plexus which gave birth to the main stem. Assuredly many branches have emerged with the parent trunk in this way, but, on the other hand, repeated instances can be given of the origin of branches from an embryonic artery after it has become an independent tube. I should assume then that the delay in the elaboration of stronger arterial coats enables the embryonic artery of more naked endothelium to respond to a stimulus and send out branches. The most fundamental example of this is furnished by the main branches of the aorta, for, although the aorta itself arises from a narrow strand of capillaries (Tiirstig), it becomes a large unbranched tube functioning for the vitelline and chorionic capillaries long before it again sprouts out branches. It is true, however, that when these branches arise, they themselves are first in the form of capillaries and often constitute a plexus — e. g., that nourishing the limb buds. All this, then, is paramount to stating that the vascular system does not grow merely at its end bed, — i. e., the capillary area — and for a time during the development of the vascular system this fact must be conceded. 6 We find, then, in the development of the embryo, that the Aeby idea of a uniform all-pervading capillary plexus anlage for the vascular system is far too crude and inexact for the facts, but that the vessels even from the beginning take definite positions and relations to the tissues, and that consequently the main vascular stems which come out of them cannot, as a matter of fact, course in every possible direction.

However, there was still a precious kernel of truth in the old idea. The vessels arise from plexuses which, if not all-pervasive, still have frequent connections with other plexuses. More important still, a functional role is played by the plexuses and the vessels supplying and draining them, and we cannot doubt but that hydro-dynamical grounds often determine which parts of an original " The ability of an embryonic artery to sprout out capillaries is, however, eventually lost, and in late fetal life, as in the adult, capillary sprouts occur almost exclusively at the peripheral or true capillary bed. In general, then, it may be held that the origin of a vessel from any of the largest arteries assigns it to a quite early embryonic appearance. This may be the underlying cause for the fact that vasa vasorum seldom arise from the vessel which they suppfy, for by the time an arterial wall becomes elaborated enough to need a proper nourishment of its own, the main vessel may have lost the power to send out direct sprouts.

It must be remembered that even the smaller arterioles and pre-capillary vessels of the adult are highly differentiated structures in which muscle and elastic elements occur so that their inability to directly sprout capillary branches is in no contrast with the possession of this power by even the largest of the embryonic arteries, for the latter structures have a greatly simplified histological structure, differing less from the capillaries themselves.

DEVELOPMENT OF THE VASCULAR SYSTEM. 585 plexus shall be converted into larger trunks. This gives the possibility of variation not only in the exact position of a single trunk but also in the territory supplied by it, for by means of its capillary union with the area of its fellow trunk it may successfully displace the latter.

Repeatedly in the history of the vascular system we find areas which are primitively supplied by many smaller vascular trunks secondarily supplied by a single large one, and this seems certainly due to the fact that the constant presence of a functioning capillary bed enables the successful artery to annex neighboring fields. Whereas in the intestine we have originally a row of vessels which go to the gut wall and yolk-sac, these later give way to three large permanent trunks (aa. coeliaca et mesenteric^), and whereas in the arm bud a row of delicate arterioles nourish the limb, soon fewer, and eventually a single artery possesses this field. This story is repeated over and over again in the vascular system from centre to periphery. It occurs first in the history of some of the main stems, as I have just indicated, but it occurs repeatedly afterwards as the more peripheral vascular tree is gradually developed. 7 All these facts now enable us to understand better many peculiarities of the adult vascular system. Above all, can we appreciate better now a reason in the frequent occurrence of vascular variations, for we see clearly the possibility for channels other than the normal ones to obtain possession of a field. Again, there sometimes occur in embryonic vessels, and more rarely in adult ones, cases of " inselbildungen " where a chief stem is for a short distance reduplicated. See, for instance, the inselbildungen at the origin of the aortic arch s in Fig. 421, or the condition of the a. hyaloidea in Fig. 430. These phenomena are difficult to explain on the basis of our old notions of the outgrowth of naked vascular stems, but appear now as cases in which an arterial stream has for a time retained two paths instead of a single one through its preceding capillary net." We cannot, however, carry this analogy further and proclaim that all instances of anastomosis and of plexiform vessels in the adult are survivals of embryonic conditions, for many of the latter are clearly secondary formations. 10 7 Witness, for example, the history of some of the arm vessels. Goppert (1909) has shown that many branches which are at one time present on the dorsal side of the chief arterial stem are later replaced by a single artery arising in their middle, the a. interossea dorsalis. (Compare his Figs. 7 and 8, Taf. viii, with Figs. 9 and 10, Taf. ix.) 8 This is doubtless due to the tendency of the aortic arches, in common with other vascular trunks, to be formed at first from true capillary vessels which are fundamentally multiple rather than single. Thus I have seen repeated instances in injections of the chick and pig where not one but two or three capillary sprouts are sent out by the dorsal aorta into one of the visceral arches, though only one of these vessels persists to constitute an aortic arch.

9 Backman (1909) has recently discussed the view here advanced.

10 It could be thought, for instance, in cases where an artery was resolved into a rete mirabile that we had a survival of the primary embryonic net here. That we may err, however, in such an interpretation is clear from the research of Tandler (1906), who showed that the retia mirabilia occurring at the base of the skull in many artiodactyls does not really represent an incomplete resolution of the primitive plexus of the a. earotis interna, but comes from a later series of capillary sprouts which arise directly from the naked carotid stem and plexify. It is, then, likely that many of the " wundernetze " which constitute the arterial channels in some mammals — e. g., edentates and pmsimians, are specialized secondary formations. The extensive subcutaneous venous plexuses of the limbs and body wall of man are also clearly secondary formations. (See beyond.)


But it is to the general conception of the developing vascular system as a connected and functioning whole, which recent studies and especially injections have given us, that a better notion of the formation of variations will accrue. The fact that the arterial current has formed its path from the capillaries, and with the shiftings of growth may form new ones through this mesh, is of the greatest significance. Thus, a chief vessel may channel a new way through the capillary paths connecting two of its branches, the old stem atrophying, and so come to acquire new relations, for instance, to neighboring nerves, as Goppert has recently proved. (See Goppert, 1909, pp. 376—379.) Vascular variations, however, do not occur in an infinite number of ways, because the developing arteries and in fact even the capillary plexuses have definite relations to the tissues. But even with these relations, the tendency to a lingering plexiform type in the main stem and the constant occurrence of the capillary mesh, any part of which may, as it were, be called into service, — all this gives sufficient choice in the selection of a permanent channel to cause the usual variations which are so frequent in the adult.

Exactly what hydrodynamical factors are concerned in the development of arteries and veins from the primary indifferent net which we have seen to exist, are not yet well known. 11 Knower's experiments certainly demonstrate that normal vascular development is dependent on the heart beat.

With our present ideas on the mechanical advantages enjoyed by a wellestablished channel, it might appear all the more remarkable that prominent embryonic vessels are not oftener retained, for example the median artery as the chief stem in the lower arm. But it is probable that continued studies on the manner of vascular development will only strengthen the conviction that the eventual dominance of secondary channels is due to the utilization of an actually better path by the blood when we consider the entire territory to be supplied. The path which the blood takes is dependent from the beginning on the demands of the tissues. Strong growth, which, so to say, sucks the blood in another direction, must play a prominent role in development, and so it may come about that a straight path is actually exchanged for a circuitous one. But this indeed is the whole course of vascular growth, for longer and longer paths are chosen by the developing arterial tree, and we are forced back to the conclusion that the growth and demands of the peripheral or capillary portion of the system exercise a determining influence on the architecture of its main stems, both in embryo, fetus, and adult.

Comparative. — In accordance with a similarity in the general anatomy of vertebrate embryos, we find also a remarkable agreement in the plan of their chief vessels. They furnish us with the opportunity of comparing accurately the vessels 1 The minor differences in the angles at which vessels arise may greatly favor or hinder their acquisition of a large part of the current in a contest of trunks supplying the same field. (Hess, 1903.) We also know, of course, that two distinct types of vessels are differentiated according as they stand in relation with the supplying or draining system, for in the former case we have always small independent thick- walled vessels and in the latter a greater number of large, anastomosing, and thin-walled ones. The measurements made by Mall and his pupils indicate that arterial blood is delivered to the capillaries in the various organs through a much smaller-calibred system than the veins must possess to drain it. Nevertheless an actual role of the circulation in adapting the architecture of the vessels needs to be investigated. (Since this was written Oppel, 1910, has published his extensive discussion of this phase of the subject, and the reader is referred to it.)

DEVELOPMENT OF THE VASCULAR SYSTEM. 587 of various vertebrates. This has been possible chiefly through the mass of splendid comparative researches which we owe to Hochstetter. The reader should consult, for this stand-point, Hochstetter's various researches and his more general pre* tat ions. He will see there that we now possess for many vessels a fundamental vertebrate plan. The first and most brilliant example of such homologies was furnished us by the work done on the homologies of the aortic arches and the vi derived from them. Rathke's (1843) work on the arches in the mammalia is a classic. Late Changes. — Finally, we may remark that the history of the development of the vascular system hardly ends with the establishment of the chief trunks, since the position of many embryonic vessels is far removed from their adult one. A remarkable shifting or wandering process must consequently take place. The studies of W. His gave us a classical example of this in the caudal displacement of the heart and of the great vessels in connection with it, and Mall first called our attention to the cervical position of the intestinal vessels which later shift into the abdomen. These great changes are usually accomplished by the time the human embryo is twenty millimetres in length and finally other, less momentous displacements occur." 2. Development op the Human Vasculak System. 13 A. The origin of the vascular system.

B. Description of the vascular system present in early human embryos.

C. The development of the arteries.

D. The development of the veins.

A. Origin of the Vascular System.

The question of the source of the cells which form the vascular system still remains, as it has for a long time, one of the most disputed problems of mammalian and indeed of general vertebrate embryology. The question has met no undisputed solution for the case of any vertebrate, and here, in contrast to the dearth of human material, we can possess a wealth of all the necessary stages. When such fundamental questions as the genetic relation between extra-embryonic and embryonic vessels, and indeed even the method of origin of the former — the well-known vitelline vascular anlagen — are still unsettled, and when we consider the paucity of these earlier stages which should be necessary for the determination of this question in man, a speedy solution of the problem in human ontogeny is expected by no one.

u Such, for instance, as that of the upper thoracic aorta on the columna vertebralis. Whereas in enibryos of 20 mm. the upper aa. intercostales find their interstitia intercostalia at the same level, in the adult, as is well known, they must course upwards to reach their interspaces.

13 When completing the present account of the development of the human vascular system, I had access to six young embryos in the possession of Professors Kollman, Eternod, R. Meyer, Strahl and Felix. These were studied in the laboratory of Professor Wiedersheim in Freiburg i. B. To all of these gentlemen I wish to express my sincere thanks. Four very valuable ernbryos in the collection of Graf Spee were studied in his institute in Kiel, for which great privilege I am deeply indebted.


If, then, we posses no safe generalizations with which to interpret the few observations possible on human embryos, we are also still further retarded by certain peculiarities of the early history of the primate embryo which affect profoundly the vascular system. The presence of an early vascularized belly stalk and chorion distorts the entire sequence of tbe usual development of the vessels and furnishes us at once in embryos astonishingly young with highly specialized and characteristic phenomena. These latter facts are now beyond doubt and I shall present them briefly below. Here only it need be remarked that the early history of the human vascular sj-stem has not enabled us as yet to make any statement as to the exact cellular origin of the endothelium in man. The only facts in the human embryo's history which may be brought into relation with this important question seem to point clearly to a mesodermal source for the primary blood-vessels. These are: 1. The abundant vascularization of the early chorion, where apparently any role of the entoderm can be excluded. 14 2. The early vitelline vascular anlagen which cause a characteristic " hummocking " of the yolk-sac wall lie in the mesodermal coat of the latter and are sharply separated from the entoderm.

3. The earliest vascular cells within the body of the embryo (in the Graf Spee embryo " Glaevecke " ) are certainly in more intimate relation with the mesodermal than with the entodermal cell layer.

B. Description of the Vascular System Present in Early Human Embryos.


It is certain that long before any vessels are present in the body of the human embryo, and at a time so early as considerably to precede the formation of any somites, typical " vascular anlagen" are found scattered over the ventral pole of the yolk-sac.

In those mammals which, like the rabbit, possess a vitelline vascular area of the limited circular form, bounded by a marginal sinus, which characterizes the lower vertebrates, it is probable that the vascular anlagen first form in a ring-like row around the borders of the future area vasculosa, as Van der Stricht has described for the rabbit. But in the primates (and presumably in all other mammals in which the yolk suffers a complete overgrowth by the area vasculosa, — e.g., carnivores and artiodactyls) the vascular anlagen are most irregularly scattered, covering at an early date the whole ventral surface and soon all the yolk-sac.

These vascular anlagen or blood islands, as in other vertebrates, appear as nodular swellings of the wall of the yolk-sac, ami consist microscopically of circumscribed cell clumps lying between the mesoderm and entoderm. The cells of these clumps very early show a differentiation into centrally lying blood-cells and a row of peripheral bounding cells, — the endothelium. The best-developed, and hence earliest, of these anlagen are situated more ventrally, the younger nearer the body of the embryo, in very "It may perhaps be mentioned that those who, like Hubrecht (190S), consider that a considerable part of the early mesoblast has really come from the entoderm will dispute the above statement.



early stages they can be shown to be concerned in the formation of the vessels in the belly stalk (aa. umbilicales), and these vessels belonging to the placental circulation, are so exaggerated in development as to precede the appearance of vessels in the embryonic body proper. Furthermore, as Eternod discovered, when, later, the vascular trunks of the embryo proper make their appearance (the aorta? and vv. umbilicales), they are already connected with the chorionic capillaries through the precocious aa. umbilicales



E. V. M.

Fig. 401. — Section of a vascular anlage in the wall of the yolk-sac in the human embryo 2 mm. long shown in Fig. 408. E., entoderm; V., vascular cells; M., usual mesoderm cells.








I I EC. E. Fig. 403.

\ V.

Fig. 402. — Section of a more advanced vascular anlage from the yolk-sac of the same embryo as shown in Fig. 401. EC endothelial cell.

Fig. 403. — Section of a well-developed vessel from the same yolk-sac.

and w. chorioplacentares, and so it comes about in the human embryo that an umbilical circulation exists in embryos so young that the mesoderm is as yet unsegniented into somites.

The embryos which I have been able to examine with respect to their early vessels constitute, together with that so fully described by Eternod, the following series, given in order of their probable age.

Designation of embryo.

Length of embryonic shield .



Von H(erff) . .37 mm. Graf Spee Graf Spee, Arch. f. Anat. u. Phy., 1896.

Frassi, NT. 1 1 . 17 mm. Prof. Keibel Frassi, Arch, f . mik. Anat., Bd. 70 u. 71.

Glaevecke. . . 1 . 54 mm. Graf Spee Graf Spee, Arch. Anat. u. Phy., 1S89, 1896.

Eternod 1.3 mm. Prof. Eternod Eternod, Anat. Anz., Bd. xv, No. 11, 12, 1898.

For further descriptions of the embryos themselves the reader is referred to Chapter TV of the present work, where this has been done by Prof. Keibel. Here we need be concerned only with remarks on what blood-vessels are present.


The embryo von Herff, the youngest, and probably only consisting of the region of the primitive streak, possesses abundant vascular anlagen scattered over the entire ventral and part of the. lateral surfaces of the yolk-sac, reaching often to the angle of junction of the yolk-sac with the embryonic shield. Whereas in general those of the anlagen whose development seems most advanced are more ventrally situated, there exist also many not widely separated from the embryonic body, in which an evident differentiation into endothelium and blood-cells has come about. In the belly stalk and chorion of this embryo there are, as Graf Spee has described, highly characteristic strands of spindle cells, which often consist of a double row of nuclei and, again, may enclose a distinct lumen. These cells appear to keep to themselves, and to constitute a single unified but widely branched tissue which grows oftenest in strands frequently anastomosing among themselves. This tendency to constitute a distinct tissue element different from the connective tissue, together with the histological appearance of longer oval nuclei and more deeply staining cytoplasm, suggests strongly that we may be dealing here with endothelium, a conviction strengthened by the typical vascular appearance given in those instances where the cells surround a distinct lumen. 15 The embryo Frassi, which now, besides the primitive streak region, shows a considerable embryonic area in front of this (the two being separated by a typical canalis neurentericus) also exhibits many evident blood-vessels. Not only have we here again an abundance of well-differentiated vascular anlagen on the ventral walls of the yolk-sac, but in many of the sections through the belly stalk vessels can be recognized (one of these being especially large), and this is also the case in the chorion.

The embryo Glaevecke of Graf Spee (NT. 2) contains, besides many yolk vascular anlagen and chorionic vessels seen in the preceding stage, the first vascular cells within the body of the embryo itself. In the region of the heart these constitute a true typical endothelial anlage for that organ (Fig 404), but also further caudalward there can be recognized many cell strands clearly isolated and different in character from the endoderm and mesoderm between which they lie; they thus occupy the typical position for, and present the typical appearance of, early vascular cells 16 (Fig. 405). At first lying about half-way between the point of insertion of the yolk-sac and the mid-line of the body, they gradually shift lateralward, and before the neurenteric canal is reached occur quite exclusively only at the 15 If such an interpretation be correct we must have a remarkable growth of the endothelium in the chorionic membrane of young human embryos, for there occurs here no coincident development of blood-cells, as is typically the case in the vitelline anlagen. These cells of the chorion (Spee's so-called spindle-cells) are present in relatively great numbers and, so far as I am aware, cannot be distinguished from similar cells in still younger embryos (e.g., that recently demonstrated by Fetzer [1910] ), where we are quite unable to distinguish vascular beginnings on the yolk-sac proper. The above suggestion, however, appears to me forced on one who now takes up the study of a series of progressively slightly older stages, such as we have in the embryos here dealt with, where difficulty would be experienced in separating these cells from those which gradually are concerned in the formation of undoubted vessels.

16 Concerning the origin of these intra-embryonic vascular cells, it can only be said that the histological appearances are inconclusive, and one may often see what might be taken for a genetic connection of the cells with the intermediate mass of the mesoderm as Mollier (1906) has reported in reptilian and avian embryos. On the other hand, however, the series of sections does not permit one to exclude the strong possibility that these cells constitute a connected unit which could have invaded the embryonic body from the splanchhopleure of the yolk-sac.



lateral margins of the embryonic area and very near the insertion of the yolk-sac. In the area behind the neurenteric canal these cells apparently retreat to the upper margin of the yolk-sac proper. This parallels the condition found in this area in the younger embryo von Herff. Finally, as the allantois is given off, these vascular anlagen can be traced into vessels which in the belly stalk lie at first on either side of and soon below the allantoic diverticulum; they are consequently in the position typical for the umbilical arteries and doubtless represent the anlagen of these vessels in the belly stalk.

We turn now to the embryo described by Eternod (1898) and in which we have the earliest circulatory conditions in the human embryo. However, a considerable gap exists between the stages which we have just been considering and that depicted by Eternod, for in the latter case a system of vessels are now present

Medullar groove

X * \


Pericardial cavity

Anlage of the cardiac endothelium

Medullar pad

Parietal blade of the mesoblast

f /.y


Pericardial cavity

Anlage of the cardiac endothelium

Fig. 404. — Cross section of the human embryo Glaevecke (collection of Graf Spee), taken just in front of the anterior intestinal portal, showing the vascular cells constituting the anlage of the endothelium of the heart. X 113.

coursing through the body of the embryo, — the aorta? and umbilical veins. Exactly how these first vessels are formed in man is as yet unknown. The umbilical veins, the heart, the aorta? and umbilical arteries, and, finally, the chorionic capillaries, form the simple vascular cycle here present (Fig. 406)." The Eternod embryo measures approximately 1.3 millimetres in length and has also as yet no indication of mesodermic somites. It shows an anteriorly placed heart, the short aortic end of which, doubtless representing the future bulbus

17 As yet, though there are many vessels on the yolk-sac, particularly on its ventral surface, no evidence for a vitelline circulation exists, for no connections between these capillaries and the aorta? can be traced. This, the most revolutionary result of Eternod's study, appears to place man unique among mammals in the ontogenetic precedence of the umbilical over the vitelline circulation, for in the mammalia generally, as is well known, the yolk-sac circulation is always primary.

592 aortaB aud aorta ventralis, gives off the aortic arch 18 which sweeps up on each side to the primitive aorta. The aortas course dorsal to the head-gut and on either side of the notochordal plate, and at last turning down sharply into the belly

Vessel-forming cells

Vessel-forming cells

Vessel-forming cells

Vessel-forming cells

Vessel-forming cells Vessel-forming cells

yolk-sac vessels

Anlage of yolk-sac vessels

Anlage of the aa. umbilicales

Anlage of the aa. umbilicales

Fig. 405. — A series of sections through the human embryo Glaevecke, showing the earliest intra-embryonal vascular cells and the anlage of the aa. umbilicales in the oelly stalk.

stalk, run out onto the chorion without having given any branches into the tissues of the embryo. The umbilical veins (w. umbilicales primitives), which collect the blood from the extensive chorionic vessels (w. chorio-placentares), unite

M It is certain that we are not dealing here with three aortic arches and that the picture given by Eternod is somewhat schematized, the small irregular vessels here likely being persisting strands of a capillary plexus. (See Fig. 398 of a duck.)



in i lie belly stalk into a single trunk (v. umbilicalis Lmpar) 3 but again pari as the embryo proper is reached, and course in the body wall on either side near the attachment of the amnion. Just as they begin their embryonic course, each umbilical vein receives a large tributary from the capillary plexus nf the yolk-sac,

Sacculua _ vitellinus

Aorta desc. sinistra

Insula Wolffi

Ansa vitellina

Opening Vena umbilicalis impar

Fig. 400. — Lateral view of the vascular sj stem in a human embryo 1.3 mm. long, without somites. (After Eternod, from Kollmann, Handatlas der Entw. d. Mensehen. 1S97 ' »g. 512.)

and these two tributaries anastomose with one another on the wall of the yolk-sac so that a venous ring is produced enclosing the allantois (ansa vitellina). The significance of this is unknown. This connection of the vitelline vessels with the umbilical vein gives the possibility of an early drainage of the yolk-sac in that Vol. II.— 38

594 direction were any aortic afferents traceable to the vitelline plexus. But at a time when such afferents clearly supply the vitelline plexus, the true vitelline veins are formed, so that at the time when we are first able to affirm the possibility of a complete vitelline circulation it is supplied and drained, as in all mammals, by its own system of vessels. 18 EMBEYOS POSSESSING FKOM SIX TO EIGHT SOMITES.

Unfortunately, we possess as yet no human embryos belonging to the interesting period in which the first five somites are formed, but for stages only shortly after this several excellently preserved and trustworthy specimens are now known. I base what can be said about the vascular system at this stage chiefly on the study of the following four embryos : 20

Designation of embryo.

Number of somites.



PfannenstielKroemer NT. 3 5-6 somites • •

[Keibel-Elze, 1908.

[Felix, 1910.

Graf Spee embryo ....

Mall No. 391 .

6-7 somites 7-8 somites

Graf Spee Prof. Mall [Graf Spee, 1887.

Kollmann, 1889, 1907 (Figs. 187 1 and 188). Dandy, 1910.

Eternod's embryo ....

8 somites (the 8th not com

pletely separated) Prof.

A. C. F. Eternod fEternod, 1896, 1899, 1904, 1909. {Kollmann, 1907 (Figs. 183 and 184)

The most striking change which has occurred in the vascular system of these embryos is found in their possession of the first branches of the aorta. The majority of these aortic branches go to the yolk-sac (aa. vitelline primitivae), and, though at present appealing as almost frank lateral branches (Figs. 407 and 444), in later stages are shifted so as to come off more ventrally.

The primitive vitelline arteries form an irregular series of connections between the aortse and the vitelline capillary plexus which has arisen out of the early yolk anlagen. They are not, as

19 Eternod has pointed out that in the Selenka specimen of Hylobates the vitelline plexus is also shown in connection with the chorionic vessels by means of two stems which surround the allantoic tube in reaching the belly stalk, and in the latter they fuse to a single large vessel (v. umbilicus impar) which can be traced into the chorionic vascular tree. The ansa vitellina may consequently be a characteristic primate structure, but it is impossible in the light of present knowledge to assign to it any significance. It is quite possible that these vessels merely represent a persisting connection of the vitelline and chorionic vessels indicating a primitive common anlage.

The manuscripts of Dandy and of Felix were generously placed at my disposal by their authors and were of much service during the study of the embryos concerned.



a rule, at first segmentally arranged. Cephalad they may extend as far as the first intersegmental cleft, as Dandy (1910) first showed. In the Pfannensteil-Kroemer embryo the first of these appear opposite the third somite on the left, and in the Eternod embryo between the third and fourth somite on the right. Occurring generally so frequently as to be opposite each somite thereafter (though they occupy no constant position with regard to the somite mass), when the unsegmented mesoderm is reached they are found in far greater numbers, and eventually resolve the termination of the aorta itself into a plexus of capillary-like vessels not unlike that to be seen in injections of chick embryos. 21 With

V. umbilicalis

Medullary thickening

Unsegmented mesoderm

V. vitellina

Fig. 407. — Cross section of a human embryo with six somites (NT. 3), in the region caudal to the last somite, showing the delicate aorta and their vitelline branches.

this plexus, as Felix (1910) has shown, and as I can confirm, the umbilical artery is in connection and so by these multiple roots takes its origin from the aorta. It is by means also of the farther caudal growth of this plexus that the aorta is continued caudally and the a. umbilicalis wanders caudalward through a considerable distance. These facts establish clearly that the umbilical artery is

21 Felix (1910) is unable to follow the aorta throughout its entire extent in the first of the above embryos (NT. 3). indicating his belief that it is not present in some places (see his account, p. 603). I am not able to agree #with his account, nor with his statement concerning the intestinal vessels (p. 606) — " cranialwarts offnen sich seine Gefasse teilweise frei in die primare Leibeshohle " !

596 merely a modified vitelline vessel; for a considerable time its roots of origin from the aorta are indistiguishable from the row of primary vitelline arteries.

Headward the vitelline plexus is connected on each side with the heart by two primitive vitelline veins, which receive the umbilical veins which have coursed in the somatopleure and then turn in sharply from their lateral position to gain the heart; in doing




Fig. 408. — Dorsa view of model of human embryo possessing 7-8 somites, being the same embryo shown in Fig. 23 B (ante, p. 32). Portion of ectoderm of right neural plate is removed, showing thickness of wall and its relation to deeper structures. The three primary cerebral vesicles are indicated. (After Dandy.) All., allantois; Ch.. chorda; Coe., ccelom; Fg., fore-gut; Hg., hind-gut; Ht., heart; Mes., mesoderm; P.c, pericardial ccelom; U ., umbilical arterial sinus; V ., umbilical vein. (Mall, No. 391 J

this they traverse the bar of mesoderm which intervenes between the pericardial cavity and the yolk-sac wall and which is destined to constitute the septum transversum of His. Their course here hence resembles entirely that taken by the terminal portions of



the primitive vitelline veins in other very early mammalian embryos {e.g., the rabbit), and 1 present here in lieu of a more detailed description a series of accurate tracings of their course (Fig. 409). Besides the vitelline circulation, which is thus well established

v. vitelloumbilicalis communis

v. vitello umbilicalis communis


v. umbilicalis v. vitellina

Heart (venous arch)

Junction of the v. umbilicalis with the v. vitellina

— v. umbilicalis v. vitellina

v. umbilicalis v. vitellina Sinus venosus

Fig. 409. — A series of sections through the human embryo with 7—8 somites (shown in Fig. 408), showing the relation of the chief venous stems to the heart.

in these embryos, other vascular channels are beginning to be formed; these constitute the first endothelial sprouts to be sent out into the tissues of the embryo proper; arising dorsally from the aorta, they lay the anlage for the a. en rot is interna in the region



of the fore- and mid-brain, whereas farther caudally they form a series of presegmental and segmental dorsal offshoots of the aorta. In all cases these tiny vessels are directed to the sides of the neural tube, which consequently, neglecting the primitive gut and yolk-sac, must be considered the first embryonic tissue to receive vessels; this occurs, in fact, before the nervous system is in the form of a tube, for it is, in the first of these embryos, a widely open furrow. Each vessel, having reached the side of the medul

Aorta dorsalis

A part of the plexus vitellinus

a. umbilicalis

Fig/]410. — Reconstruction of the arterial system of a human embryo with 6 somites (NT. 3), seen from the left. (Modified after W. Felix, 1910.)

lary furrow, divides T-like and can be traced a short distance caudally and cephalically. It is by the anastomosis of these branches that in older embryos (those of 15 somites) a longitudinal vessel is established at the sides of the hind-brain and the neural tube caudal to it {v. capitis medialis). The dorsal segmental arteries have long been known, for they occupy accurately the interspaces between the somite masses ; some four of them are already present in the Eternod embryo with 8 somites, and, since these are progressively smaller in size cephalo-caudally, their outgrowth from the aorta quite certainly proceeds in this sequence.




Designation of embryo.

Number of somites.



Bulle, NT. 5 Pfannenstiel III, NT. 6 Graf Spee No. 52 13-14 somites 14 somites 15 somites Prof. Kollmann Graf Spee

Kollmann, 1889, 1907. KeibelandElze, 1908 Low, Jour. Anat. and Phys., 1908. Felix, 1910.

__ .

In human embryos which possess some fifteen somites we not only have an increased number of dorsal segmental arteries (eleven

Cauda) most part of otic thickening

v. card. ant.

v. card. ant.

v. vitello umbilicalis communis

v. vitello umbilicalis communis

v. card. ant.

v. vitello; umbilicalis communis d. v. card, ant,

Fig. 411. — A series of cross sections through a human embryo with 15 somites (collection of Graf Spee, No. 52), showing the course and relations of the primitive head vein (v. capitis medialis et v. cardinalis anterior) and the relation of the chief venous stems to the heart.

in the embryo with fifteen somites) and of the primitive vitelline arteries, but also another set of aortic branches which I shall designate as the primitive lateral branches of the aorta. 22 These 22 Graf e (1905) was, I believe, the first to see these vessels, describing them in the posterior portion of a chick embryo of about sixty hours; but their cephalic extension was shown by Williams (1910), who described them in the first two intersegmental clefts.

600 vessels take origin from the lateral aortic wall, often from its ventro-lateral angle, and course obliquely upward and outward in the space between the somite and the intermediate mesodermic mass; here I have seen them anastomose with the cardinal vein; they are also often connected with the dorsal segmental arteries by direct cross anastomoses in the loose mesoderm of the intersomitic clefts.

a. umbilicalis

Aorta dorsalis

Fig. 412. — Reconstruction of the arterial system of a human embryo with 14 somites (NT. 6), seen from the left. (Slightly modified, after W. Felix, 1910.) The primitive vitelline arteries in the area opposite the first five somites have atrophied, but the series of these vessels begins from here caudally to form a continuous row unrelated apparently to metamerism and finally, in the unsegmented area, giving way, as in the younger embryos, to a plexus from which the umbilical artery takes origin.

The first venous channels of the embryo proper — the w. cardinales anteriores — are found at this stage, and in the Graf

DEVELOPMENT OF THE VASCULAR SYSTEM. 601 von Spee embryo with 15 somites can be traced clearly from the region of the optic vesicles, cephalically, to their opening into the common vitelline and umbilical vein, caudal ly. These veins, as Grosser (1907) has shown to be probable for all vertebrates, in man also possess two different and distinct topographical relations, for in their cephalic course they lie close to the sides of the neural tube, constituting the v. capitis medialis, whereas more caudally — i.e., in the region beginning with the first mesodermic somite — they take up a lateral position between the somite and the ccelomic mesoderm where thev may be designated the true vv. eardinales

Right horn of the sinus venosus

Ventral wall of the yolk-sac'

Pericardial cavity Cut edge of the septum transversum

Fig. 413.— Sinus venosus and septum transversum in a human embryo with 14 somites (NT. 6), viewed from above. The right half of the septum has been removed to show the sinus venosus contained therein. (Drawn from the model by Dr. Alex. Low.) anteriores. Opposite the third somite the vessel finally joins the common vitello-umbilical vein by coursing dorsal to the ccelomic cavity in this region (Fig. 411). The v. capitis medialis receives several (four) direct dorsal offshoots from the aorta, so that it really appears as a longitudinal neural anastomosis of these presegmental dorsal arteries ; it turns out rather sharply somewhat in front of the first somite to constitute the true anterior cardinal, which is again formed apparently by a laterally-situated longitudinal anastomosis of loops formed by the dorsal segmental vessels; to it also the primitive lateral branches of the aorta are joined. 23 23 This history of the formation of the anterior cardinal vein in man is thus identical with that which has been previously found in the chick (Evans, 1909), in which latter embryo Williams has recently described the same phenomena in a careful account of the region about the second somite. It is consequently probably significant that the picture furnished by the endothelial cells constituting the first dorsal segmental vessel in the Eternod embryo shows a marked lateral wandering of the endothelium (Fig. 439). This probably should be accounted the first staee in the formation of the anterior cardinal in man.

602 The vitelline veins still behave in all essentials as in the younger stages ; they receive the umbilical veins when quite lateral in position, and the common veins receive the anterior cardinals, turning in sharply to constitute the sinus reuniens (Fig. 411).


In stages which are intermediate between those which have just been described and embryos possessing limb buds, the posterior cardinal veins develop. This has already occurred in the embryo with twenty-three somites (Eobert Meyer, 300, N.T. 7),

Anastomosis extending along the neural tube

a. umbilicalis

Region of the'om phalomesenteric artery

Fig. 414. — Reconstruction of the arterial system of a human embryo with 23 somites (NT. 7). (Aftei W. Felix, 1910.) *which has no indication as yet of limbs. It is probable that lateral loops of the dorsal segmental arteries are instrumental in the formaton of these veins, as is the case with the anterior cardinals. 24 At this stage the dorsal segmental vessels form in the

24 This method of formation of the posterior cardinal veins appears fundamental. Raffaele (1892) and Hoffman (1893) described it for selachian embryos and Grafe (1905) and the writer have indicated it in the case of the chick.



tissue of the intersomitic clefts large well-marked vascular arches or loops, one limb of which is against the neural tube while the other joins the cardinal vein (Fig. 436). At this stage also the primitive lateral branches of the aorta form an extensive system, and at many levels we are able to find all three systems of branches occurring together and segmentally arranged.

The row of vitelline arteries is by no means exclusively segmentally arranged; nevertheless there is a symmetrical disposition in that these vessels occur in pairs, ^o that in the region in

9th somite

V. cardinalis post.

V. urnbilicalis

X " —v

i .""i^ 3r*wf'f5 I

4 c


I* " V - • c » '.a -"'V 1 !



V. um "bilicalis


v (.

i V. omphalomesenterica

  • -^.>

Fig. 415. — Cross section of the human embryo with 23 somites, shown in Fig. 414, taken through the region of the 9th somite, showing paired aa. vitelline.

which the two aortae primitive have fused to a single median aorta we can observe that two vessels arise from the ventral surface of the aorta and course each on its corresponding side of the gut, which possesses as yet no mesentery (Fig. 415). When later an intestinal mesentery is formed, these vessels course for a time side by side, but eventually are completely fused to a medial ventral trunk, or it is possible that one member of the pair gains the ascendency and its fellow atrophies.


In the posterior region of the body we find not only the posterior cardinal vein, but a new one, lying ventral to the former and near the ccelomic epithelium. This vein, the v. subcardinalis (F. T. Lewis, 1902), has probably arisen by sprouts from the posterior cardinal trunk, as Graefe has shown for the chick; at any rate the presence of a large number of anastomoses between these two vessels speaks strongly for this view. In the region of the mesonephros the subcardinal vein occupies a characteristic position ventral to the )Yolffian duct, but at levels above this region, where we have as yet, according to the view of Felix, only the pronephric anlage, the vein is also found, and in the same position, i.e., ventral to the chief duct, as Fig. 416 will show. There also

12th somite V. card. post.

Principal duct Pronephros Chamber

V. subcardinalis $-fc

Epithelium of the coelom

Fig. 416. — Section showing the position of the v. cardinalis posterior and the v. subcardinalis in a human embryo with 23 somites (NT. 7), taken in the region of the 12th somite.

occurs in embryos of this age another vein medial to the pronephros, and it has probably arisen as a longitudinal anastomosis binding together vascular offshoots from the posterior cardinal veins and also the primitive lateral branches of the aorta. 25 EMBRYO OF 4.9 MM. LENGTH (35 SOMITES, N.T. 14).

By the time the embryo reaches a length of 5 millimetres several important changes in the vascular system have occurred. The embryo described by Ingalls (1907) and shown in Figs. 417, 418 will serve to illustrate this stage.

It will be noted that four complete aortic arches are present, and that another pair — the sixth or pulmonary arches — are being 25 One finds in this embryo pictures which very much resemble that given by Grafe in his table 11, Fig. 7, for a chick of 71 hours.

DEVELOPMENT OF THE VASCULAR SYSTEM. 605 formed by both ventral and dorsal endothelial sprouts. The third pair, the carotid arches, are by far the largest of the series, while the first are already very much reduced. The aortic root on each side now appears to continue toward the head beyond the location of the first arches. This, the internal carotid artery, is doubtless the trunk representing the very early capillary sprouts which the first arch sent toward the brain. It courses headward lateral to the hypophysis and bending dorsally anastomoses with a long branch — the a. vertebralis cerebralis — given off from the first of* the dorsal segmental arteries here present (in this case the hypoglossus artery). At the optic cup the internal carotid gives off the a. ophthalmica as its first branch, and somewhat beyond this a very large branch (a. cerebri ant. et med.) which courses forward between eye and brain; other smaller branches are given off to the mid-brain region, and the carotids then sweep backward to join the cerebral vertebrals, which they furnish with their main volume of blood, although later the stem of origin of the latter vessel gives it most of its blood. The first cervical dorsal segmental artery 26 anastomoses with the hypoglossus, and consequently the path is already furnished for this stem to take over the cerebral vertebral when the hypoglossus yields the current and atrophies. Excluding the hypoglossus vessel twenty-seven dorsal segmental branches arise in pairs from the aorta and sacralis media artery, — i.e., the full number of cervical, thoracic, and lumbar vessels and the first two sacral segmentals. The umbilical arteries, though they later shift to the last lumbar level, arise here opposite the third lumbar segmentals, the remaining lumbar arteries at this stage consequently arising from the a. sacralis media. The aa. umbilicales course medial to the Wolffian ducts, but at the prominent bend which they make in turning upward are in connection with capillaries lying lateral to the Wolffian duct, which ultimately gain a connection with the aortic wall and completely displace the medial roots of origin of these arteries. The subclavian artery arises from the seventh dorsal segmental pair.

Most interesting are the ventral branches of the aorta, for these no longer form a uniform row of vessels, but reflect a beginning differentiation of the gut. Consequently there are retained, besides many smaller ones, three chief branches, to correspond to the stomach-pancreas region, the vitelline-duct region, and the colon respectively (a. cceliaca, a. omplialomesent., et a. mesent. inf.). The middle of these three stems, which is also by far the largest, since it drains yolk-sac as well as gut, takes origin from 20 1 refer to the segmental artery cranial to Hochstetter's "first cervical artery," naming that artery the first cervical which courses with the first cervical nei've, as do Mall. Tandler, and Broman.




A.s. A.c.



A . omphal

Fig. 417. — Reconstruction of the arterial system in a human embryo 4.9 mm. long, lateral view. After Ingalls, Arch. f. mik. Anat., Bd. 70, p. 530, 1907.) A.c, a. cceliaca; A.c. a., a. cerebralis ant.; A.c.s., a. caudalis sin.; Ag., optic vesicle; Al., allantois; A.m.i., a. mes.inf.; A. o., a. ophthalmica; A. omphal., a. omphalomesenterica (with three roots); A. p., a. pulmonalis; A.s., a. subclavia; A.u.s., a. umb. sin.; A. v., a.vertebralis; B. 1,2,8,4,6, aortic arches; C.a.l, first cervical artery; Cd., caudal intestine; D.p., dorsal pancreas; D.i., ductus vitello-intestinalis; Gb., gall-bladder; Ha. , hypoglossus artery; lb., inselbildung; La., lunganlage; M., stomach; Oes., oesophagus; T. a., truncus arteriosus; UK., lower jaw; V., questionable union between the a. vertebralis and a. car. int. (N. T. 14. ) the aorta by four distinct roots. It will be noticed that all these branches are much above their location in the adult, as can be seen by comparing them with the dorsal segmentals opposite, and it is not indeed until the embryo attains a length of from 16 to 20 millimetres that their definitive position is reached.



V. c. a.

D. C.

V. c. p.


Fig. 418. — Reconstruction of the venous system of the embryo shown in Fig. 417. A., fibres of the accessorius; C.l, first cervical segment; D.C.8., ductus Cuvieri sin.; F., facialis; G., glossopharyngeus; L.l, first lumbar segment; 0., ear vesicle; 0.1, first occipital segment; S-l, first sacral segment; T., trigeminus; T.l, first thoracic segment; V., union of the left umbilical vein with the liver circulation; V.c.a., v. card, ant.; V.c. p., v. card, post.; V. i., v. ischiadica; V.u.s., v. umb. sin.; V.u. 8.*, remains of the original circulation to the sinus venosus; X., linguo-facial vein.

Irregularly arising, lateral branches of the aorta go to the mesonephros.

In contrast to the earliest stages, the venous system of the embryo proper is now well developed, and one sees the well


known fundamental pattern of the two cardinal veins on each side uniting to form the ductus Cuvieri which then joins the umbilical. The anterior cardinal can be seen beginning in two strong efferents in the head region, the first of which doubtless represents the ophthalmic vein. Passing medial to the ganglion of the fifth nerve, the main vein next receives a tributary from the hypophysis region, and continues caudally on the lateral side of the acusticofacial ganglion, the auditory vesicle, and the ganglion of the glossopharyngeus, but medial to the vagus ganglion ; just before reaching the latter nerve, it receives a prominent tributary from the dorsal region (v. cerebri post.. Mall), although smaller tributaries have joined it all along its previous course. Before joining the ductus Cuvieri, several venules run into it, which, from their position and correspondence with the veins of other mammalian embryos, can be recognized as the segmental veins belonging to the first cervical and the several occipital segments. The ductus Cuvieri receives on each side a slender venule, which drains the capillary plexus in the first visceral arch. This vessel crosses from the latter • into the ventral body wall (here constituted by the membrana reuniens over the front of the heart) and runs in this to open into the ductus (v. linguo- facialis, F. T. Lewis, 1909).

The posterior cardinal vein begins at about the level of the third lumbar segment, and courses in the tissue dorsolateral to the Wolffian body. It receives the dorsal segmental veins, which do not become appreciable structures until about the level of the fifth thoracic segment. The seventh cervical segmental vein receives the vessel from the arm bud (the subclavian vein), although this afterward shifts up to the anterior cardinal vein. The v. subcard inalis can first be recognized at about the level of the seventh thoracic somite, and empties into the posterior cardinals at the level of the sixth cervical one. They drain the Wolffian body at this stage, but later acquire greater significance inasmuch as they are incorporated in the formation of the inferior vena cava. The umbilical veins are already sending their main mass of blood into the liver, but with their old connections with the sinus venosus still evident. This uppermost and superficially lying part of the umbilical vein receives tributaries from the arm buds, and this source of blood delays their atrophy (Evans, 1909). The vessels from the lateral body wall also drain into the v. umbilicalis along all of its course until the liver is reached, so that the vein forms at this time an important drainage channel for the entire lateral body wall. The vitelline veins empty their blood directly into the liver sinusoids, the blood from the left omphalomesenteric vein being collected by a short trunk which enters the left horn of the sinus venosus (r. hep. sinistra) ; but the right and larger one possesses a wide passage through this organ to the right horn of the



sinus venosus. The two vitelline veins are anastomosed on the ventral, then on the dorsal, and again on the ventral sides of the duodenum, forming thus two venous rings around the gut (His).


The vascular system present in an embryo measuring 7 millimetres begins to be complex enough to demand detailed descriptions for many areas, so that with a brief presentation of the chief features here, we may leave this account of the early vascular system as a whole and turn to the explicit history of the various vessels. The main blood-vessels in an embryo of this length (7 mm.) are well known to us through the papers of Mall (1891),

A. bas.


2. AB.


A. v. c

5, AB.

4. AB.

"Nff. AB.

A. car. ext.

A. p.

Fig. 419. — Profile reconstruction showing the arterial system of the head and neck in a human embryo 7 mm. long. (N.T. 28.) fAfter Elze, Anat, Hefte, Bd. 35, Heft 106, Taf. 16, Fig. 3.) S, S, 4, 6, and 6 AB, aortic arches; A. bas., a. basilaris; A. car. ext., a. carotis externa; A. p., a. pulmonalis; A.v.c, a. vertebralis cerebralis; Obi., ear vesicle.

Piper (1900), and Elze (1907) ; the latter 's figures are here reproduced and his description largely followed (Figs. 419 and 420). Three complete aortic arches exist, the third, fourth, and sixth. The first pair, already very weak in the preceding embryo (4.9 mm.), are now entirely atrophied, but both dorsal and ventral end pieces of the second arch are recognizable, the dorsal remnant, in fact, being destined to constitute the trunk of origin of the stapedial artery (Tandler, 1902). The upper end of the sixth arches sends out, between these and the fourth pair, a small vascular frag Vol. II.— 39

610 ment, which perhaps represents the persisting upper end of a previous transitory fifth arch. The aorta ventralis in the area of the first and second arches now constitutes the a. carotis ex

cap, i.

29. SA

V. subc.


A.i. 26. S A.

Fig. 420. — Profile reconstruction of the same embryo, showing general arterial and venous system. (After Elze, Taf. 15, Fig 2.) X 26.5. A. coe„ a. cceliaca; A. i., a. iliaca; A. s., a. subclavia; A., a. omphalomesenterica; Atr. s., atrium sinistrum; A. u. s., a. umbilicalis sin.; D. At., ductus venosus Arantii; V. card, a., v. cardinalis anterior; V.cap.l., v. capitis lateralis; V.extr., extremity vein; V. hep.r.s., v. hepatica revehens sinistra; V.o.m., v. omphalomesenterica;, v. omenti minoris; V.mes.s., v. mesenterica superior; V.u.s., v. umbilicalis sin.; V.subc, v. subcardinalis; D.C'uv.s., ductus Cuvieri sin.; 7, 20, 25, 29 SA., segmental arteries; Sin. ven., sinus venosus.

terna which reaches into the region of the upper jaw. The internal carotid artery, after giving off the dorsal rudiment of the second arch, courses headward to give off in the region of the hypophysis

DEVELOPMENT OF THE VASCULAR SYSTEM. 611 a branch which courses toward the brain beneath the optic cup, then above this, a small branch to the latter (a. ophthalmica), and dorsal to the eye, a large branch which apparently supplies the main portion of the fore- and 'tween-brain (a. cerebri ant. et med.). After giving off other branches to the lateral side of the 'tweenand mid-brain, the artery ends in its ramus communicans posterior which appears to continue the main trunk into the basilar artery. The a. vertebralis cerebralis now arises from the first cervical segmental artery, and the preceding hypoglossus vessel has entirely disappeared. The stem of the first cervical vessel, however, has been shifted cranially until it is now opposite the sixth aortic arches, whereas earlier even the hypoglossus artery arose relatively further caudad. The two cerebral vertebral vessels unite in the mid-ventral plane to form the a. basilaris, which extends from the area opposite the vagus nerve to the vicinity of the oculomotor nerve, where it splits into the two posterior communicating rami which connect it with the internal carotids. Both the cerebral and the basilar arteries send off many branches to the hind-brain, some of the branches of the former anastomosing dorsal to the emerging fascicles of the twelfth nerve, so that these fibres appear to go through arterial fenestras. The full number of cervical, thoracic, and lumbar segmental arteries exist and all but the last sacral. The umbilical arteries come off the aorta at the level of the last lumbar segments, and now course lateral to the Wolffian ducts and send each a small branch into the posterior limb (a. ischiadica). The ventral branches of the aorta have been reduced to three main trunks: the a. cceliaca arises opposite the fourth thoracic artery; the a. omphalomesenterica is two-rooted, its upper root coming from a level slightly above the fifth thoracic artery, its lower one opposite the sixth; while the a. mesenterica inferior arises opposite the second lumbar vessel. 27 The veins show several important changes. The proportionately great growth of the head gives a great drainage area for the anterior cardinal vein, which is consequently much increased in size. Its foremost tributary is the paired primitive sinus sagittalis superior on the top of the fore-brain. These are the only tributaries to reach the mid-dorsal plane, with the exception of a very short partly paired one on the roof of the mid-brain. 28 Sev "A ventral vessel is seen arising from the sacralis media beyond the level of the fourth sacral segmental vessels and supplying the end of the gut where it goes over into the cloaca. Branches of this type from the sacralis media are seen in injections of other mammalian embryos, namely the pig, where they are more numerous and reach further headward.

18 Considerable interest should attach to these mid-dorsal veins of the midbrain, inasmuch as Grosser (1901, p. 322) has demonstrated a pair of veins in this locality in bat embryos where they are retained, in fact, in the adult as the v. longi


eral tributaries from the first two visceral arches reach the main trunk, while behind the ear vesicle the posterior cerebral (Mall) vein has grown to great proportions. The anterior cardinal, proceeding caudally, surrounds the vagus by a venous ring and then goes under the hypoglossus to join the posterior cardinal trunk. The ling no -facial vein (F. T. Lewis) is now no longer a tributary of the ductus Cuvieri, but joins the cardinalis anterior.

The posterior cardinal vein, receiving the same number of dorsal segmental tributaries as is sent out from the aorta and sacralis media artery, drains also the marginal vein of the leg bud and along its entire length receives efferents from the Wolffian body; but the axillary vein now reaches the ductus Cuvieri and will soon indeed be a branch of the anterior cardinal trunk. The subcardinal veins (not illustrated) exist from the level of the tenth thoracic segment caudally and are in frequent communication with the corresponding posterior cardinal. The umbilical veins can no longer be traced to the ductus Cuvieri on either side, and the superficial portion of the primitive vein is only represented by several small stems draining from the body wall into the main trunk just before it plunges into the liver. The left umbilical is by far the larger of the two, and the same is true for the vitelline trunks; the right vitelline indeed has atrophied practically completely, and its previous large pathway through the liver has given way to many sinusoidal paths, whose supplying or portal stem may be called the ramus arcuatus while the corresponding draining or hepatic venule is the v. hepatica dextra. The v. hepatica sinistra still opens independently into the sinus venosus. The main mass of the umbilical blood takes a direct path through the liver in the ductus venosus (Arantii), which receives several small veins from the caudal end of the stomach (Magenvene, Hochstetter, 1893; v. omenti minoris, Broman, 1903). From the liver capillaries a slender vessel grows out into the caval mesentery, the anlage of the v. cava inferior.

tudinales meseneephali, apparently homologous with this vein in reptiles (Grosser and Brezina, 1895) . Grosser attributes the disappearance of these veins in other rnammals to the great overgrowth of the cerebral hemispheres, which, as is well known, are notably small in the Chiroptera. He also calls attention to the fact that Salvi (1897) probably saw the same structure, if we may judge from the descriptions in his paper on the development of the meninges in Cavia and Lepus. Attention may here be called to the fact that the primary head capillary plexus in pig embryos halts in its spread along two parallel mid-dorsal lines in the mid-brain as well as fore-brain, and, just as the medial margin of the former plexus comes to constitute the primitive paired sinus sagittalis superior, so also in some embryos an exactly similar condition transitorily exists on the top of the mid-brain. So that from the careful description oil Grosser and the less satisfactory references of Salvi this evidence may now be added to indicate quite clearly, I think, that a condition resembling the reptilian v. longitndinalis meseneephali exists transitorily in all mammalian embryos, including man.



Since the human embryo, like that of all other vertebrates, possesses a row of definite gill bars or visceral arches, separated distinctly, externally by clefts, internally by entodermal pockets or pouches, so also its primitive vascular system is in conformity with this fundamental plan, and strong branches connecting the dorsal and ventral aortae — the aortic arches — each course in a visceral arch (Fig. 421). It has been known for a long time that in all vertebrates above the fishes, — i.e., in the amphibia, the sauropsida, and the mammalia — the number of these arches is five A.d. J.

Fig. 421. — Model of the pharynx and aortic arches in a human embryo 5 mm. long. (After Tandler, Morph. Jahrb., xxx, 1902, Taf. v, Fig. 17.) A. d., aorta dorsalia; C. a., conus arteriosus; J., island; IV, fourth aortic arch; VI, sixth aortic arch.

Within the last three decades, however, it has gradually been shown that in reality six arches exist in these classes, the fifth aortic arch being everywhere an exceedingly transitory vessel. 29 20 In 1886 Van Bemmelen for the first time described a transitory aortic arch found between the systemic and pulmonic arches in embryos of birds and reptiles. A year later Boas (1887), in welcoming this discovery, pointed out its importance in the comparative anatomy of vertebrates. He recalled (Morph. Jahrb., Bd. 7, 1882: Bd. 8, 18S3) that in amphibian larvae, as well as in Ceratodus, Polypterus, and Amia, the four aortic art-lies which occur occupy the third to the sixth visceral arches, the pulmonic artery being given off in each ease by the last pair, i.e., by the aortic arch of the sixth gill-bar. It hence appeared evident that the pulmonic arch was the sixth and not the fifth of the series in all vertebrates, and Boas now predicted the discovery of a transitory fifth arch in the embryos of mammalia, the only remaining class in which it had not as yet been seen. Two years later Zimmermann, as is well known, reported the presence of a fifth arch in embryos of man, the rabbit, and the sheep. Tandler's careful paper, following in 1902, reported traces of the arch in the rat and two very clear examples of it in man, for which he furnished the first figures published. Lehman has described what were


Ziimnemiann (1889) was the first to indicate that there was any tendency to the formation of a fifth arch in man, reporting the separation of the fourth arch into two distinct vessels in a seven millimetre human embryo.

In his article on the development of the head arteries in mammals, Tandler (1902) described two very clear cases of a human fifth aortic arch, neither of which, it may be noted, corresponded to Zimmerrnann 's description, for in both cases the fifth arch took origin from the aorta ventralis and joined the dorsal portion of the pulmonary arch. A diverticulum of the fourth endodermal pouch (postbranchial body) separated the fifth and sixth arches, whereas the fourth pouch lay between the fifth vessel and the fourth arch. Since then other observers (Elze, 1907) have reported the partial presence of this vessel in the same situation. The question of the existence of a true fifth aortic arch was soon seen to involve the identification of the postbranchial body as the fifth branchial pouch. Hammar (1904), now, had described an embryo of 5 mm. (N.T. 20), in which five pouches were present, the fifth {using with the ectoderm of a fifth branchial cleft in the manner typical for these structures. Elze (1907), aware of Hammar 's report, and finding a fifth ectodermal cleft opposite the post-branchial body in an embryo of interpreted as vestiges of the arch in the rabbit and gave a distinct instance of it in the pig. It is important that in some of these cases — e.g., in Tandler's (man) and in Lehman's (pig) — we had also to do with what was apparently a fifth pharyngeal pouch. This structure, the so-called postbranchial body, is not new. It had been known to appear behind the fourth pouch but soon to grow into the latter, with which it had a common opening into the pharynx and of which it now appeared to be a diverticulum. It seemed highly significant also that in the cases which have just been enumerated the fourth pouch pointed towards the ectoderm between the fourth and fifth arches, while the postbranchial body occurred between the fifth arch and the sixth. The whole picture of these interrelations, in short, pointed strongly to their all being serial true branchial structures. In 1906 F. T. Lewis • indicated that the very general acceptance of this interpretation was probably caused by the weight of comparative considerations. He called attention to the ordinary conception of a vessel which could be called an aortic arch having a definite course from the ventral to the dorsal aorta, and emphasized that a fifth arch of this completeness had never been seen, except by Zimmerrnann in the rabbit, where subsequent investigators (Lehman, Lewis) have not been able to confirm him. This fact must be admitted, for even in Tandler's clear cases the accessory arch does not join the dorsal aorta, but instead fuses with the pulmonary arch before this ends dorsally. Locy has emphasized that it seems generally true that the fifth arch is in some way connected with the last pair, in some of the lower classes, in fact, the pulmonic arch appears to have split, — e.g., Lacerta (Peter). Other reports have recently come in affirming fifth aortic arches in other mammals, — Soulie and Bonne (1908) in the mole, Coulter (1910) for the cat, and Reinke for the pig, — (Note on the Presence of the Fifth Aortic Arch in a 6 mm. Pig Embryo, Anat. Record, vol. 4. No. 12, December, 1910) and the evidence is too unanimous to cause doubt that vascular rudiments in the position of a fifth arch occur generally in the mammalia.



7 mm. (N.T. 28), felt no hesitancy in identifying the postbranchial or ultimobranchial body as the fifth branchial pouch. Finally Tandler (1909) has examined a considerable number of embryos bearing on this point and brought together all that has been ascertained about the fifth arch. His conclusions seem to put the question at rest and tc show that in man, very transitorily, in embryos from five to ten millimetres in length, a true fifth arch exists (Figs. 422 and 423, A and B), springing from the truncus aorticus just before the fourth arteries are given off, and coursing dorsally in what is sometimes a distinct fifth gill bar to open into the sixth arch close

4th endodermal pouch, ventral diverticulum

_4th aortic arch


4th endodermal ' pouch, dorsal diverticulum

5th aortic ^ arch

^ ;V I - 1 **K/

jmpT 5th endodermal / pouch ^ ^ >


1 y

Arteria pulmonalis^

(Jth aorti c arch

Larynx anlage CEsop iiagus

Fig. 422. — Model of the pharynx and aortic arches of a human embryo 7 mm. long. (After Tandler, 1909.) ( Embryo H> of the I anat. Lehrkanzel, Vienna.) to its upper end. In relation with it is a special transitory branch of the vagus nerve (ramus posttrematicus, Elze), 30 in front of it is the fourth entodermal pouch, and behind it the postbranchial body (fifth pouch). The latter is indeed in early stages apparently only a caudal ventral division of the fourth pouch. It is later

30 Ramus posttrematicus der IV Schlundtasche," "E. posttrematicus V," or " R, posttrematicus II des Vagus," given off by the N. laryngeus superior shortly after its departure from the ganglion nodosum and described by Elze (1907) and Tandler (1909) in embryos from 7 to 9 mm. in length. It is usually absent in later stages, although Grosser (1910) believes he has identified it in an embryo 19% mm. (crown-rump), passing through the foramen thyreoideum.

616 4th endodermal pouch

3d 3d endo- 4th aortic dermal aortic arch fpouch arch

5th 5th endo- 6th aortic aortic dermal arch arch pouch Fig. 423 A.

3d aortic arch

3d endodermal pouch

4 th aortic arch

4th endodermal pouch

5th aortic arch 5th endodermal pouch 6th aortic arch

Fig. 423 B.

Fia.423Aand B. — Model of the pharynx and aortic arches of a human embryo 9 mm. long (XT. 37).

(After Tandler, 1909.)



incorporated in the thyroid gland (Tandler, 1909, Grosser, 1910), although apparently not contributing true thyroid tissue (Grosser).

The only certain facts which have been established in the metamorphosis of the human arches into the trunks of the permanent vascular system have been incorporated in the diagram of Fig. 424. As far as their actual arch portions are concerned, the first two aortic arches are commonly lost, but the third and left fourth arches are retained, becoming the root portion of the internal carotid and the arcus aortae respectively. On the other hand, both the ventral and dorsal aortae beyond the position of the third arches are preserved, the former to furnish the stem of the external carotid, the latter the second part of the internal carotid

A. subclavia

A. car. int. A. car. ext.

A. subcl. Aa. pulmonales A. subclavia

A. car. ext.

^A. car. int. -A. stapedialis

Aortic arch

A. subclavia

Truncus pulmonalis

Fig. 424. — Diagram of the aortic arches and their fate in man.

artery; whereas the ventral aorta between the third and fourth arches becomes the stem of the a. carotis communis. The corresponding part of the dorsal aorta disappears, so that now all of the internal carotid blood courses by way of the ventral stem. The sixth arch is lost on either side beyond the origin of the corresponding pulmonary artery, but on the right its proximal portion, between the truncus and the a. pul. dextra, persists and is the root portion of the adult right pulmonary artery. On the left side, however, this proximal portion of the pulmonic arch is apparently incorporated as part of the truncus pulmonis, and the adult a. pul. sinistra consequently is merely the exact analogue of the embryonic vessel (Bremer). 31

31 This subject forms one of the few instances in which a correction is necessary in the conception originally given us by Rathke in his epoch-making monograph on the " Aortenwurzeln und die von ihnen ausgehenden Arterien der Saurier." Rathke, as is well known, represented the right and left pulmonary arteries both coming each from its respective arch in lizards and birds, but for the snakes among


This, then, is the general outcome of the arches, although we are now in the possession of some facts concerning the fate of the first two arches about which nothing hitherto has been known.

Before proceeding to consider details of the changes undergone by the various arches, mention may be made now of several kinds of shifting or growth displacements which affect these vessels and which make it easier to understand the relations which characterize the chief trunks derived from them in the adult. In the first place, as His clearly showed, the place of insertion of the aortic truncus into the anterior pharyngeal wall, whence it is split up into the arches, moves gradually lower down, so that, while at first the arches go off horizontally and even more caudally placed from the truncus, they soon course in an ascending direction from

the reptilia and for the mammalia he showed the two vessels arising from only a single arch, in the snakes the right one and in the mammalia the left one. Rathke's ideas were founded on appearances given by embryos which have passed the earliest stages of origin of the pulmonary artery. His first showed that the earliest human puhnonaries came each from its respective arch, as in the lacertilia, and Bremer has proved that this is a general fact for all the mammalia, and suggested the high probability of its general occurrence, at least in earlier stages, in all air-breathing vertebrates. Bremer's studies have included man, the rabbit, cai, dog, pig, opossum, sheep, guinea-pig, cow, and deer, and, as a result of them, he distinguishes two methods for the formation of the adult pulmonary stem in mammals. In the method occurring most generally (man, cat, dog, rabbit, sheep, cow, deer, and opossum) the original symmetry is disturbed by an absorption of the proximal part of the left arch into the truncus pulmonalis, so that the left pulmonary, artery now rises from the bifurcation place into left and right arches while the right artery comes off its usual distance from this bifurcation place. With the destruction of the distal part of the right arch up to the point of origin of the a. pul. dextra and the eventual atrophy of the corresponding part of the left arch, the adult plan is reached, and this therefore means that we must consider the left pulmonary artery as representing only the original embryonic one, but the right pulmonary vessel has also as its most proximal part the right pulmonic arch. A second method followed in the evolution of the adult mammalian pulmonaries is exemplified by the pig and guinea-pig, in both of which forms the two early pulmonary arteries are joined in a general capillary plexus, the anastomosis enabling one root to serve as a common stem, which in the pig happens to be the left original artery and in the guinea-pig the right. Consequently, in the former animal the blood to both lungs must first traverse the proximal parts of the left arch and left original artery, and in the latter animal the corresponding parts on the right side. Sakurai (1904) has declared that the left artery in the deer moves toward the right past the bifurcation of the truncus pulmonalis to the right arch, but Bremer questions his interpretation, and the case here must rest till an examination of more abundant material in the stages implicated. In the meanwhile, Bremer's work has shown the incorrectness of the conventional diagram in which both definitive pulmonaries are shown as sharing equally the proximal parts of the sixth arches, for in no mammal is this true. Man and most of the other members of the class have a right a. pulmonalis which is of this nature, but an a. pul. sinistra, which is merely the original pulmonary artery of that side, the corresponding proximal part of its arch having been assimilated in the truncus pulmonalis.



the caudally placed root stem. These changes have been described as a " moving down " of the insertion place of the truncus, and are doubtless due to the same phenomena of unequal growth which cause the apparent rapid descent of the heart from its earliest position at the end of the fore-gut. This change takes place in a regular and characteristic way, as Figs. 425, 426, 427 clearly indicate. Originally, when only two arches exist, the truncus may

V. carrli- — 14 « nalis ant. Iv^l

V. urnbilicalis

Sinus reuniens _Sinus reuniens

V. umbilicalis

Vv. omphalomesenteric-;!'

Vv. omphalomesentericse.

Fios. 425 and 426. — Reconstruction of the aortic arches in two human embryos, measuring 2.15 mm. and 3.2 mm. respectively. (After W. His, Anat. mensch. Embry., iii, Leipzig, 1885, p. 1S6, Fig. 119, and Atlas iii, Taf. ix, Fig, 12 and 15.) be described as splitting to send on either side of the gut an ascending and descending limb — the first and second arches respectively. Soon the full complement of arches is present, and the downward progression of the aortic truncus with respect to the gill bars now gives a different arrangement of its arches from the parent stem. Both of the first two arches arise together from an ascending stem, while the third arch courses back practically horizontally from the truncus and the last two come off together from a descending

620 stem. Next an exaggeration of the length and importance of the common ascending stem for the first two arches (the stem which will later constitute the external carotid trunk) occurs and a truly ascending course for the third arch, although the latter has not yet been incorporated in the larger ascending trunk (Fig. 426). In the next changes which take place, the third arch has been carried up in the general ascending trunk (now the common carotid trunk),

V. cardinalis ant.

V. cardinalis post.

V. umbilicali;

Vv. omphalomesenteric^ Fig. 427. — Reconstruction of the aortic arches in a human embryo 4.2 mm. long. (See Figs. 425 and 426.) whereas the fourth tends to course more nearly horizontally, Eventually even the fourth and sixth arches come to have an ascending course.

At the same time that these changes in the arrangement of the arches have been taking place, another of a more general nature has transpired, for not only the heart but the whole system

DEVELOPMENT OF THE VASCULAR SYSTEM. 621 of arches also has moved down toward the thorax. A reliable criterion of this general dislocation is furnished by the relation of the arches to the dorsal segmental arteries, for the latter have a fixed relation to the somites of the dorsal body wall. Before the stage of five millimetres, all the series of dorsal segmental arteries, including the hypogiossus artery, are considerably below the junction place of the sixth arch with the dorsal aorta. By the stage of seven millimetres this place corresponds to the first cervical dorsal segmental, by the stage of nine millimetres to the second vessel, and by the time the embryo has reached eleven and a half millimetres to the sixth or even the seventh cervical segmental, from which trunk the subclavian and vertebral arteries arise (Tandler). This relation is at last almost that of the adult, where the subclavian comes off the transverse portion of the aortic arch.

At the stage of seven millimetres, a splitting of the truncus begins, proceeding from above downward and separating the fourth arches, with the system lying above them, from the sixth ones. The latter then come to have an independent common trunk, — the truncus pulmonalis, — and this, as is well known, is exclusively connected with the right heart, whereas the truncus aorticus is similarly in relation with the left.

Still another growth change in the arrangement of these vessels is to be mentioned. We left the last three arches in a markedly ascending course. Such a course obtains for the pulmonic arches so long as they persist, but after the division of the truncus the systemic truncus elongates much, pushing, as it were, the proximal portions of the third and fourth arches again upward and giving them a horizontal or even slightly descending course (Tandler).

The dorsal part of the right fourth arch now atrophies beyond the origin of the subclavian stem, and this whole segment now constitutes but a branch of the persisting a. anonyma.

A. Carotis Interna and its Branches. — It has already been emphasized that the earliest branch of any of the arches consists in that given off by the dorsal part of the first arch toward the embryonic mid-brain. This persists and is of increasing importance, and when the atrophy of the connecting portion of the dorsal aorta between the third and fourth arches results, it constitutes, together with this part of the dorsal aorta and third arch, the internal carotid artery. The internal carotid, then, consists of three morphologically different portions, — a proximal or root portion derived from the third arch, an intermediate portion consisting of the original aorta dorsalis from here to the first arch, and an end portion which is the earliest branch of the first arches and is the chief supply of the brain. 32

32 Injections of very early bird and mammalian embryos show that the trunk of the internal carotid arterv extending from the first arch distalward is


It is to be noted that, besides the larger internal carotid which is given off from the end of the first arch, the aorta dorsalis also sends several smaller branches toward the hind-brain before the region of the primitive segments is reached, and, when, at length, the latter territory is reached, the dorsal segmental vessels. Those dorsal branches which are in front of the segmental area are very transitory, and attract onr interest chiefly because they represent the first vascular sprouts sent out by the dorsal aorta into the tissues of the embryo in this region and, directed toward the sides of the medullary tube, are directly responsible for the formation of the v. capitis medialis. 33 represented at first by the outgrowth of a plexus of capillaries from that arch (Figs. 393, 394). This plexus spreads over the sides of the early mid-brain first, then over the fore and hind-brains (Fig. 395). Soon out of the several capillary stems of origin from the first aortic arch, one is chosen to become the artery and the remainder perish. Gradually the plexus of capillaries invades the ventral surface of the brain and tends to halt there on either side of a narrow mid-ventral non-vascular strip. In the meanwhile the continuation of the main arterial stem is being evolved out of this plexus in such a way that the carotid, after giving off an a. ophthalmica, appears to have two terminal branches anterior and posterior. The latter connects up with the medial ventral margin of the capillary mesh on each side, and so there come to be in the midventral region two long parallel vessels, the continuation of the posterior terminal branches of the two carotids. This conversion of the medial margins of the ventral capillary mesh here is analogous to the formation of the aorta? from the medial margins of the vitelline capillary plexus. It will be shown that in the spread of the capillary plexus over the spinal cord an exactly similar phenomenon takes place, — that is, the capillaries halt along two parallel lines on either side of the midventral plane. In the cord region also these two plexus margins are converted into two transitory longitudinal arteries, furnished at every segmental point by blood from the segmental arteries and connected headward with the same vessels supplied by the carotid arteries. The whole structure from head to tail has been called by Sterzi the tractus arteriosus primitivus; by De Vriese the primitive anterior spinal arteries. It will be evident from all this that this primitive midventral vessel is the earliest arterial anastomosis between the carotids and dorsal segmental vessels. The second of the dorsal segmentals is the hypoglossal artery, and that part of the anastomosis between it and the carotid is of the greatest importance, for it, according to De Vriese, is the a. vertebralis cerebralis of His; at any rate it is destined to form the basilar artery in the region beneath the hind-brain. Thus the basilar artery is primitively paired and gets its chief supply of blood from the carotids, for the hypoglossal artery cannot figure greatly. Gradually the double basilar is replaced by a single vessel, which is really formed through the development of anastomoses between the two parallel trunks permitting the original left vessel to persist in some areas and the right one in others. This is also what happens as regards the anterior spinal artery. Very soon after the unpaired basilar is produced, its lower source of blood exceeds its upper in importance, and when the cerebral vertebrals are taken over by the cervical vertebrals, the latter vessels are the main supply of the basilar.

83 These pre-segmental branches of the aorta have, of course, another interest, inasmuch as we may be dealing with evidences of a segmentation of the head in front of the occipital somites. Be this as it may, Tandler (1902) has seen a remarkable row of these vessels in the rat. where they seem to arise at regular intervals.

DEVELOPMENT OF THE VASCULAR SYSTEM. 623 As soon as the region of the somites is reached the dorsal aortic branches are strictly segmentally arranged, — i.e., they course between successive somites. The pair between the first and second somites, however, early atrophy, and the pair situated between the second and third somites and which are in relation with the hypoglossus nerve remain somewhat longer and, as the so-called hypoglossus arteries, constitute the first of the series. In embryos of five mm. length (Tandler 1902, Ingalls 1907) the hypoglossus can be seen giving off a long longitudinal cranial-coursing branch, which headward anastomoses with the a. carotis interna on each side, thus making two long arterial arches. This branch of the hypoglossus artery is the a. vertebralis cerebralis. Later, as has been mentioned, the a. vertebralis cerebralis is taken over by the first cervical segmental artery, and the hypoglossal artery atrophies, and still later, as was first shown by Hochstetter (1890), an anastomosis between the first seven cervical segmentals (aa. vertebrates cervicales) enables the seventh of these vessels to act as the origin for the vertebral artery. De Vriese has pointed out that in all early embryos the carotid, after giving off the ophthalmic artery, may be considered as dividing into two terminal branches, anterior and posterior, the latter of which turns round to anastomose with the a. vertebralis cerebralis and is by far the more important of the two. When the cerebral vertebrals fuse to a basilar artery beneath the hind-brain, the two posterior terminal branches of the carotids consequently join each other in this trunk. This is the condition of the arteries in the head in embryos measuring nine millimetres (Fig. 428). Here the ophthalmic artery is not illustrated, but the carotid is seen splitting into its two terminal trunks, a small anterior and a strong posterior, the latter continued into the basilar. The anterior terminal trunk immediately gives off the anterior chorioidal artery and proceeds as a prominent vessel on the side of the fore-brain, encircling the optic cup from above and meeting its fellow of the opposite side just behind the olfactory pit. This vessel is the a. cerebri anterior, and gives off many rami to the cerebral vesicle, which are later represented by a single (Compare his Fig. 8, p. 302.) De Vriese (1905) mentions their appearance in the rabbit (see her Fig. 28, planehe 16), and for the area in front of the hypoglossus vessel mentions two as being more constant, one at the level of the otic vesicle and the other near the gasserian ganglion. In the human embryo KroemerPfannenstiel (N.T. 3), with six somites, I have found two of these vessels on the left in front of the first somite, and in the Eternod embryo, with eight somites, one behind the region of the first aortic arch, just in front of the second pharyngeal outpocketing ; whereas in the Spee embryo No. 52 there are on each side, although not paired, four of these pre-segmental dorsal branches of the aorta. Finally, Ingalls in his 4.9 mm. embryo distinguished clearly four of these vessels on the left side. The relation of these vessels to the cranial nerves or visceral arches awaits demonstration.

624 trunk, the middle cerebral. The posterior terminal branch of the carotid gives off many branches to the sides of the mid-brain, and these later are also represented by a single trunk, the posterior cerebral. In the next succeeding stages we see an increase in the

A. chorioidealis ant.

A. cerebri ant. et med

A. vertebralis

Fig. 428. — Graphic reconstruction of the arterial system in the brain of a human embryo 9 mm. long. (After Mall, Amer. Jour. Anat., vol. iv, Plate I, Fig. 4.) (Mall No. 163. ) importance of the anterior chorioidal artery (Fig. 429), but it is remarkable that single large stems representing either the middle or posterior cerebral artery are very late in appearing. Mall is

Vena cerebral.

post.Qater sinus petros. sup.)

Sinus transversus

— Oonfluens sinuum

Sinus sagit. inf.

Sinus sagit. sup.

A. chorioidealis ant.

cerebri ant.

i~S/& lHI&\ ^^ Carotis interna. v ~^££§HiP^ Vena

Fig. 429. — Graphic reconstruction of the vessels of the brain in a human embryo 38 mm. long. (From Kollmann, after Mall.) (Mall No. 145.)

of the opinion that we must consider the last-mentioned artery as being represented originally by all the small branches which come off from the carotid between the third and fourth nerves behind and the middle cerebral in front. In older embryos (48 mm. long) these

DEVELOPMENT OF THE VASCULAR SYSTEM. 625 many branches are represented by a large mesencephalic artery and a small true posterior cerebral (Mall) ; in older fetuses the latter branch absorbs the former.

The ophthalmic artery is the first branch of the internal carotid to develop. In embryos measuring seven millimetres it can be seen to course toward the eye, dividing in its mid course into the a. ciliaris longa temporalis and a common trunk, afterwards splitting into the a. ciliaris longa nasalis and the a.hyaloidea. The latter artery pierces the optic cup, courses through the vitreous body, and reaches the posterior surface of the lens in capillaries. The arrangement and size of these branches of the ophthalmic are such that the a. ciliaris longa temporalis appears as the continuation of the main stem, and this is true up to the stage of 20 millimetres at least. The ciliary arteries supply a capillary plexus representing the chorioidea. Dedekind (1908) has reconstructed this simple vascular scheme in an embryo measuring 19 millimetres (Fig. 430 and 431). The hyaloid artery is noted by Dedekind as turning into an arterial plexus before being resolved into the capillaries constituting the tunica vasculosa lentis. Here, then, is another instance of several paths being used by the arterial blood before the reduction to a single path. The hyaloid artery serves as the later a. centralis retina, but no retinal vessels are present till late. The researches of 0. Schulze (1892) had indicated the same fact in other mammals. Versari (1903) has stated, indeed, that the human embryo reaches 120 millimetres in length before the retinal vessels are formed. In an embryo of 33.4 millimetres Dedekind has recorded the a. lachrymalis, aa. ethmoidales, and a. nasofrontalis.

We have as yet only an incomplete record of the development of the eye vessels in man, but Versari has furnished important observations on older stages (beginning with 22 mm.). In the splendid paper by Schultze the older stages in many mammals were beautifully portrayed, and some of the eye vessels in human fetuses of the sixth and eighth months shown. However, only Fuchs's careful study in the rabbit can lay any claim to completeness.

Fate of the Second Aortic Arches. — As a rule, no trace of the first arch is seen in embryos of 7 millimetres and only the dorsal and ventral ends of the second arch are evident. Tandler (1902) has recently declared that in man and other mammalian embryos the dorsal parts of the second arches become the root portions of the stapedial artery on each side. 34 The a. stapedialis persists throughout life in some mammals, — e.g., the rat, — but normally atrophies in man. At the height of

34 Although the recognition of an embryonic artery piercing the mammalian stapes dates back some thirty years (Salensky, 1880), no one had before established the relation of this vessel to the aortic arches.

Vol. II.— 40

626 A. ciliaris longa temporalis

Nervus opticus

Inner lamella of optic cup Outer lamella of optic cup

A. ciliaris longa nasalis A. hyaloidea Point of entrance of the a. hyaloidea in the optic nerve Vasa hyaloidea propria Tunica vasculosa lentis"

Fig. 430. — Left eye of a human embryo 19 mm. long, opened through a horizontal section. X 66. (After Dedekind, 1908.)


Anlage of the chorioidea


A. ciliaris longa nasalis

V. ophthalmica superior

Nervus optu

A. ophal mica it

Bulbus Fig, 431. — Left eye of the same embryo seen from the temporal side. X 66. (After Dedekind, 1908.)



its development it possesses, after piercing the anlage of the stapes, three branches, which follow the three divisions of the fifth nerve ; these are the supra-orbital, the infra-orbital, and the mandibular rami, respectively. The first of these (ramus supra-orbitalis) leaves the main stem, shortly after the stapes is passed, so that the infraorbital and lower-jaw rami have a common stem (Fig. 432). The infra-orbital division of this stem passes behind the third division of the fifth nerve to gain the second division, which it follows. Later (in embryos of 15 to 17 mm.) the external carotid artery anastomoses with the common trunk for the infra-orbital and mandibular rami, just at the point where these vessels are given off. The infra-orbital ramus gains the outer side of the third branch

Fig. 432. — Profile reconstruction of the head vessels and nerves in a human embryo 12.5 mm. long. (After Tandler, Morph. Jahrb., xxx, Taf. v, Fig. 21.) R. s., R. i., R. to., ramus supra-orbitalis, infraorbitalis, and mandibulars of the a. stapedia; L., a. lingualis of the a. car. ext.

of the fifth nerve by the development of an arterial loop around the nerve and the atrophy of the medial limb of the loop. Soon the original common trunk of the infra-orbital and mandibular rami (which lies above the point of the anastomosis witli the external carotid) becomes surrounded by the auriculo-temporal nerve and we can recognize in it the future a. menmgea media. Xow the stapedial atrophies from its origin to its division place into the three rami, and consequently these branches are then all supplied by the a. carotis externa, the stem of supply for the supra-orbital branch being the old common stem for the two lower branches, in which the flow is now reversed ; this is, as has been said, the middle meningeal artery, whereas the ramus infra-orbitalis is the a. infra

628 orbitalis of the internal maxillary, and the ramus mandibularis, the a. alveolaris inferior. This is clear from the diagrams in Fig. 433.

The place of origin of the stapedial artery and its relation to the stapes identify it accurately with the second visceral arch, but its territory of supply, when its three typical rami are developed, is entirely in the province of the first arch. This becomes intelligible when we know that these rami are later acquisitions of the stapedial, that primarily they arose from the first arch, and were later added to the a. stapedialis. Such, at any rate, is the case in the rat, as Tandler was able to show that the blood supply of the jaws (upper and lower) came originally from the dorsal part, of the first arch. To the stem supplying the jaws, a supra-orbital vessel was added, and then from the stapedial vessel an anastomosis with this common stem developed, whereby the three branches went



R. m

A. m. m.

V. a. t.

A. c. i.

a I) c Fig. 433. — Schemata showing the fate of the a. stapedialis in the human embryo. (After Tandler, 1902.) a represents the conditions present in a human embryo 17 mm. long, b those in one 19 mm. long, and c those in one 23 mm. long. II., second branch of the trigeminus; III. , third branch of the trigeminus; A.m.m., a. meningea media; A.c.c, a. carotis communis; A.c.e., a. carotis externa; A.c.i., a. carotis interna; N.a. t., nervus auriculotemporal; R.i,, ramus infra-orbitalis; R.m., ramus mandibularis; R.s,, ramus supra-orbitalis.

to the a. stapedialis. This early history of the three stapedial branches has not as yet been secured in man, but the facts at present known make it none the less certain that the stapedial artery here has gained the territory of the first arch only secondarily. In man the three branches of the stapedial, instead of being derived from the dorsal end of the first arch, are probably derivatives of the ventral portion of that arch and the aorta ventralis. 38

33 It has been known since the time of Rathke that in many adult reptiles an artery exists which pierces the columella. The same reptiles possess another artery which supplies the upper jaw and courses with the vidian nerve. It would be of the greatest interest if ernbryologieal observations here should establish the origin of the vidian -aceomjimnying artery from the first aortic arch and the columella-piercing vessel from the second arch, like the stapedial of mammals. Evidence that this may be true is furnished by an interesting variation found by Grosser (1901) in a young mammalian embryo (bat). Here the infra-orbital branch of the stapedial artery was not a member of the usual trunk, but an independent branch of the carotis interna, having a definite relation to the vidian nerve, just median to which it coursed. If these homologies, which were suggested by Tandler, are established, then



A. Cakotis Externa. — The trunk of this vessel may be considered the aorta ventralis from the origin of the third arches cranialward. His indicated that the lingual artery was among the first of its important branches to develop, and at 17 millimetres (N.T. 65) Tandler identified the superior thyroid, lingual, and external maxillary arteries. These vessels are, in fact, present at 14 millimetres, when the internal maxillary is also being evolved from the anastomosis of its trunk of origin with the stapedial (Fig. 434). At this stage one also sees a prominent branch of the carotis externa coursing dorsalward. This is the a. occipitalis, having the

A. occipitalis \

A carotis ext. v

A. carotis com.

""• A. lingualis

Fig. 434.

A. thyreoidea % superior \ A. maxillaris ext.

-Graphic reconstruction of the face vessels in a human embryo measuring 14 mm.

Mall collection.)

(No. 144,

position and typical relations of this vessel to the muscle masses. Its proportionatel} 7 great development in these early stages is probably to be explained by its importance as a meningeal vessel.

1. The territory of the first aortic arches in all the higher vertebrates is supplied at first by vessels corning from that arch. The stem for these vessels or one of them may course with the vidian nerve.

2. The territory of the second arch possesses a vessel normally related to the columella (or stapes).

3. The second vessel (stapedial) remains in its original state in the reptiles mentioned, but in the mammals usually annexes the branches developed from the first arch.

4. In adult mammals the stapedial artery secondarily surrenders these branches to the external carotid and atrophies, or, in the cases where is persists, at least loses its mandibular ramus to the external carotid artery (rat), and in some cases also its infraorbital one (bat).

630 At 15.5 millimetres, the chief superficial branches of the carotis externa are evident, the a. auricularis posterior and a. temporalis super ficialis.

Nothing is known of the development of the coronary arteries. Tandler has noted their beginnings in a 17 mm. embryo (N.T. 65).

The only observations known to nie (1901) on this subject are the fragmentary ones of Martin (1S94) and those of F. T. Lewis (1904). Lewis has called attention to the fact that the heart of early embryos is nourished by diverticula of the ventricular lumen which course between the muscular trabecular — sinusoids of Minot, the chief method of nourishment of the myocardium in the lower vertebrates. Later the coronary system supervened and there was a great regression of the extensive sinusoidal system characteristic for the preceding stages. Lewis records the coronary arteries being first recognizable hi rabbits of 14 days and 18 hours.

Variations. — The variations in the great vessels arising from the aortic arch have been known for a long time and could be explained satisfactorily on an

Fig. 435. — Reconstruction of the lung anlagen and their vessels in a human embryo 10.5 mm. long.

(After His, 1887.)

embryological basis ever since the work of Rathke. They have been classified by Krause, for instance, and by so many, following him, that it will not be necessary to consider them here. De Vriese's work has shown the morphological character of the posterior communicating artery, — i.e., this vessel represents the original caudal continuation of the posterior terminal branch of the carotid. Consequently, cases in which the posterior cerebral arteries appear to be supplied by strong posterior communicating vessels, represent merely a retention of normal embryonic conditions, whereas the complete atrophy of the posterior communicating is an exaggeration of normal development. Islands in the course of the basilar are readily intelligible from the original paired nature of this vessel.

Comparative. — In the fish, amphibia, birds, and reptiles the internal carotid arteries are the sole source of supply for the brain, or nearly so, since the vertebrals are unimportant. The carotid in these classes divides into its anterior and posterior terminal branches, and the latter are continuous down the spinal cord with the anterior spinal artery, baring formed the basilar in the region of the hind-brain. This is the simple scheme represented in early mammalian embryos.

The development of the main vessels in the early lung is known to us from the observations of His (1S87). His showed

DEVELOPMENT OF THE VASCULAR SYSTEM. 631 that the two pulmonary arteries are from the first asymmetrical, in that the right vessel passes in front of the so-called eparterial bronchus, whereas the remainder of its course, like the entire extent of the opposite artery (a. pulmonalis sinistra), is behind the bronchial tree (Fig. 435). The pulmonary veins, on the other hand, are placed ventral to the bronchial system, and this relation persists throughout life, giving us arteries separated everywhere from veins by the corresponding divisions of the bronchial tree. 30 Flint (1906) has followed the developing vessels in the lung of the pig, more completely than has been done in the case of any other mammal. The pulmonary veins are reported by most observers as growing out of the sinus venosus before the development of the pulmonary arteries (see also Federow, 1910). In this connection, Flint has suggested that the early appearance of a drainage channel ventral to the pulmonary anlage and the ventral projection of the anlage from the walls of the foregut combine to favor the mechanical establishment of arterial paths dorsal to the organ. These early relations are only repeated in growth, and hence may be regarded as fundamental in determining the architectural interrelations of bronchial and vascular trees in the adult organ. In relation with this is the fact that the eparterial bronchus receives a ventrally placed arterial supply, and that here, consequently, the veins and arteries are accompanying vessels. It seems hardly necessary to refute the error of Aeby (1880) and others who attempted to make the arrangement of the arteries responsible for the form of the bronchial tree. As Flint has emphasized, the arteries are mere passive followers of the bronchi in development, and arise secondarily from the capillary mesh which enveloped a newly formed diverticulum of the bronchus. 36 THE BRANCHES OF THE AOETA.

As has been seen from the preceding description, the history of the development of the arterial system in the human embryo shows that at first two long channels exist — the descending aortae — which course through the entire length of the embryonic body and emerge in the belly stalk without having sent off any branches into the tissues of the embryo. The aortae and their system of branches, then, do not develop like many other vessels of the body, but pursue an elongated unbranched course over an area into which later they are destined to send out a copious supply of arteries. When, as development proceeds, capillaries are finally sent into the embryonic tissues, these sprout from the aorta, dorsally at strictly inter- segmental points, often ventrally and laterally also at such points, but in the case of these vessels usually more irregularly.

The segmental position is strictly observed only in the case of the dorsal branches. These from the first course only in the planes between the primitive segments. The ventral branches, however, are often found arising at more fre 89 Since the above went to press I note that Pensa has given us reconstructions of the pulmonary arteries in two human embryos, 11.5 and 25 mm. long respectively. Antonio Pensa, " Osservazioni sulla morfologia e sullo sviluppo della ai'teria pulmonalis nell' uomo." Boll, della Soc. Med. Chir. di. Pavia Comunicazione fatta nella seduta del 8 Aprile, 1910. Pavia, 1910.

632 quent intervals from the aortic wall, while the lateral branches, except the earliest stages, depart furtherest from a segmental alignment. Both ventral and lateral branches, however, show a tendency to adhere to the segmental plan. 87 Recent investigations on mammals and birds indicate that the branches supplying the limb arise from the aorta at multiple irregular points as a typical capillary plexus (see beyond), but are later segmentally arranged, as is the case in the earliest stages yet seen in man.

The aortic branches fall into three groups or rows, a dorsal row, a lateral row, and a ventral row. At first the dorsal segmentals supply only the central nervous system (the spinal cord and its ganglia), the lateral row, only the Wolffian body, and the ventral row, only the primitive intestine. 38 But of these branches, those which are at first purely neural in their area of distri

Longitudinal anastomoses along the neural tube

v. card. post.

v. subcard. lat

v. subcard.'* medial.

Longitudinal anastomoses along the neural tube

v. card. post.

ramus seg. dors.

ramus lat.

ramus ventralis aorte.

Fig. 436. — Reconstruction to show the branches of the aorta in a human embryo with 23 somites (NT. 7). The reconstruction was made from six successive sections in the mid-thoracic region.

bution come eventually to supply also the body wall with its muscles and skin, and those at first purely nephric to supply also the gut branches which persist,

the adrenals and the sex glands

,T I am aware that Broman, for instance, bases much of bis discussion of the position of the ventral branches and their changes on the supposition of their being primarily segmentally arranged. This, however, is not the case, as my experience with embryos of from six to twenty-three somites clearly proves. Many of the ventral branches are unquestionably as far as possible from a segmental alignment, so that the most which can be said here is that a segmental influence is evident, but expresses itself imperfectly. Later, however, there is a marked agreement with the segmental plan, so that we have conditions analogous to what occurs in the limb buds where stages of a more irregular row of primitive limb arteries are succeeded by those in which these vessels are segmentally arranged.

38 Felix (1910), chiefly on comparative grounds, assigns the primitive function of the intestinal arteries to the supply of the pronephros. There seems, however, little evidence for this in human ontogeny, where these arteries are from the first truly intestinal vessels and where the pronephric rudiments are not in relation with these but with the primitive lateral branches of the aorta.



however, supply, as they do in the embryo, the alimentary tract, the organs derived from it (liver, pancreas), and the spleen.

It is interesting to note that Mackay (1889) constructed a hypothetical schema classifying the branches of the aorta in a similar way, some twenty years ago. The main features of Mackay's classification are thus substantiated by development, for, though he confused some secondary with the primary characters of these vessels, he recognized that there were three kinds of them, naming them, from the influence of adult anatomy, the parietal, the intermediate, and the visceral branches.

The ventral branches arise first, owing to early importance of the vitelline circulation, the dorsal branches quickly after them, and, after an interval, the lateral branches. Although Eternod (1898) did not find any of these branches in his embryo of 1.3 mm.

Fig. 437, — Cross sections of injected chick embryos showing the development of the dorsal segmental vessels. A, cross section of a chick of 50 hours (24 somites), showing the loth dorsal segmental vessels; B, a chick 60 hours old; C, 78 hours old; and D, 116 hours old: all in the neighborhood of the 20th segmental vessels. S.A., dorsal segmental artery; P, C, posterior cardinal vein; S.V., dorsal segmental vein; S. C.P., spinal ganglion's capillary plexus; R.B., ventral radicular branch of the segmental artery; S., first extra-myotomal or skin branch of the segmental artery; A. C, a. centralis; S. P., superficial capillaries without the myotome; /., probable intercostal artery.

length, many of the ventral branches and two of the dorsal series occur in embryos with six somites (N.T. 3), while in an embryo with thirteen somites (N.T. 6) many distinct lateral branches can also be recognized. Both ventral and dorsal branches grow out before the primitive aorta? fuse, and consequently when this occurs an accurate apposition of the two aorta? permits these branches to come off in pairs from the single aorta descendens.

Dorsal Segmentals (Neural Segmentals, " Segmental Arteries" (of many authors), Interprotovertebral Arteries (P. Albrecht), etc.). — The dorsal segmental branches of the aorta have often been referred to as the parietal or body wall segmentals, and, inasmuch as they furnish the large well-known intercostal and lumbar arteries, their segmental nature is preserved and recognizable

634 in the adult. These later branches of the dorsal segmentals (i.e., aa. intercostales et lumbales) so far outstrip the primary trunks in growth that in the adult they themselves become known as the branches of the aorta, and the original dorsal segmentals merely

Ramus cutaneus dorsalis medialis Ramus cutaneus dorsalis lateralis

Ramus dorsalis of the anterior branch of the a. intercostalis

Ramus posterior medialis Ramus spinalis Ramus anterior canalis spinalis A. spinalis anterior A. intercostalis

Fig. 438. — Diagram of the behavior of a typical dorsal segmental artery in the human adult. (Founded on Toldt, Spalteholz, Sterzi, and Grosser.)

Fig. 439. — The first dorsal segmental artery in a human embryo with S somites. (Collection of Professor Eternod, videp. 594.) The endothelium is seen growing in the loose tissue of the first intersegmental cleft.

as their posterior branches (rami posteriores). The course of development, however, shows clearly that the reverse is actually the case.

The dorsal segmentals begin to grow out from the aorta at about the time that the embryo possesses six somites (Fig. 410). The number of dorsal segmental arteries increases rapidly, and in



embryos in which the extremities are recognizable, almost the whole series is present. The first pair of these vessels between the first and second somite early atrophies, although they are still clearly evident in embryos of 14 and 15 somites (N.T. 7 and embryo Graf SpeeNo. 52). 30 The second pair constitute the vessels which are known as the hypoglossus arteries. These remain in embryos of five mm. in length, but shortly thereafter also atrophy, so that the first cervical pair — i.e., the arteries between the third and fourth somite, which course with the nn. cervicales 1 — are next the first of the series. As Hochstetter long ago showed for the rabbit, and as is evident for man from the Normentaf el of Keibel and Elze, the whole upper six of the cervical dorsal segmentals atrophy and the seventh only is permanent as the trunk of origin of the vertebral and subclavian arteries; this also functions as the root of origin for the eighth cervical and first (or first and second) thoracic arteries by its strong a. inter xo stalls suprema, so that the next permanent dorsal segmental behind the seventh cervical is the second or third thoracic one.

The following table shows the number of dorsal segmental arteries present in several young embryos.

Designation of embryo.

Number of somites.

No. of dorsal segmental arteries.

Probable identity of the dorsal segmental arteries.

Pfannenstiel-Kroemer, NT. 3. . .

Eternod Pfannenstiel III, NT. 6 Graf SpeeNo. 52 Rob. Meyer 300, NT. 7 Broman, NT. 11 G. 31, NT. 14 Chr. 1, NT. 28 6 8 13-14 15 23 ca30 35 40 2 4 6 11 21 23 29 29 Oi, 2 .

Gi, O2J Gi, G 2 Gi, Gjj Gi — G4.

Oi, O2; Ci-CsJ Ti.

', Ci— C$', Ti— T12; Li.

— O2; Ci-Csj T1-T12; Li, L2.

— O2; Ci— Cs; Tr— T12J Li— L5; Si— S3. 5 Ci— C8,' Ti— T12,' L1-L5; Si— S«.

In their simplest form the dorsal segmental arteries consist of single capillary loops which extend from the aortas to the venae cardinales posteriores (Fig. 437, A), yet numerous other capillaries soon sprout out from these loops; and the aortic end of the original capillary loop becomes the dorsal segmental artery and the venous end the dorsal segmental vein.

Inasmuch as the dorsal segmental arteries constitute at first the arterial supply of the spinal cord, their history belongs to that of the blood supply of the cord.

29 The first pair of the dorsal segmental arteries is not generally referred to. Hochstetter (1903), for instance, states that the first pair of these arteries courses with the hypoglossus nerve, as a result of the embryos which he, Zimmermann (1S90), and Piper (1900) had studied. These embryos were so old that the first pair of the segmental arteries had already atrophied.



With the exception of the brief account by His (1886), this subject has not been followed in detail in man; on the other hand, the main facts in the history have been ascertained for the birds (chick) and the mammalia (sheep, pig) by a series of injections, and the brief description given is based mainly on these.

The single capillary loops which constitute the early dorsal segmentals approach the spinal cord near its ventrolateral angle and the ventral part of its lateral surface. In succeeding stages these loops give off delicate sprouts, which reach the cord at the area mentioned and anastomose with corresponding capillary sprouts given off by the adjacent segmentals, thus forming a longitudinal chain of

Fig. 440. — Successive stages in the development of the anterior spinal artery in the pig. The embryos were injected and the cord dissected in the region of the first three thoracic segments. A, an embryo 8.5 mm. long, B 9 mm. long, C 14 mm. long, D 15.5 mm. long, and E 28 mm. long.

capillaries on the lower lateral surfaces of the cord. These capillaries soon increase, growing over the spinal ganglia and forming a close plexus over the lower lateral surfaces of the cord, which extends dorsally as far as the under edges of the ganglia and their roots. Ventrally this plexus extends to the ventrolateral margin of the cord. Along the latter line sprouts begin to grow ventrally, and the earliest and more important of these, occurring near the chief trunks of the dorsal segmentals, represent the future aa. radiculares ventrales. As yet no capillaries have extended beyond the myotomes. Such are the conditions which occur in mammalian and human embryos until a body length of six or seven millimetres is reached. In the succeeding stages the blood stream in the segmental artery em

DEVELOPMENT OF THE VASCULAR SYSTEM. 637 phasizes in each case two main branches out of the many capillaries, an upper or dorsal and a lower or ventral branch. The upper branch courses just ventral to the spinal ganglion and the dorsal nerve roots, joining the general plexus that more intimately invests the cord just ventral to the line of emergence of the dorsal roots, — a. radicularis dorsalis; the lower branch courses ventral to the ventral roots, extending on to the ventral surface of the cord, — a. radicularis ventralis. In the next changes which occur the most striking feature is the behavior of the capillaries on the ventral surface of the cord. The. plexus which had previously begun to extend there advances from both margins until a line is reached on each side corresponding to the lateral limits of the bodenplatte; along this line they halt temporarily in their spread, thus producing a peculiar and highly characteristic vascular pattern which leaves the middle third of the ventral surface — beneath the bodenplatte — devoid of vessels but its outer thirds covered with a close net. The medial margins of this net are soon somewhat enlarged, constituting two parallel longitudinal vessels, the primitive anterior spinal arteries (tr actus arteriosi primitivi). Very soon delicate transverse capillary bridges cross the middle area which was previously non-vascular (Fig. 440). Some capillary sprouts arising from these primitive anterior spinal arteries push into the substance of the cord and course dorsally, ending usually within the gray matter of the ventral horns. These are the future aa. sulci (Adamkiewicz), or aa. centrales. This stage of double anterior spinal arteries was first seen in the human embryo by His (1886). It is probably most definite and typical for human and mammalian embryos from 9 to 11 mm. in length. His's observations showed it well marked in the human embryo of 10.9 mm. and still apparent in one of 13.8 mm.

The anterior radicular arteries contribute directly to the anterior spinal on each side, and the latter vessel is really to be viewed as merely a particularly prominent anastomosis between these aa. radicales ventrales. In like manner, in later stages, a strong arterial anastomosis develops between the posterior radicular arteries and is known as the posterior spinal artery.

To return now to the general development of the dorsal segmental vessels and their system of branches, we find, at the stage which we are considering, these vessels each possess two chief branches, the anterior and posterior radicular arteries, which are concerned respectively in the formation of the longitudinally coursing anterior and posterior spinal arteries, and which as development proceeds become separated more and more from the cord itself by the formation of the meninges, which (in the adult) they must pierce before reaching the cord.

But besides these two branches of the dorsal segmentals, another soon develops which sprouts out beyond into the skin. This is the representative of the trunk which later gives off both the muscular and cutaneous rami; the former do not as yet exist, so that the vessel may be said to be the ramus cutaneus dorsalis medialis (ramus posterior medialis of Grosser, Fig. 438). Below this another branch of the dorsal segmental now extends out ventral to the anlage of the rib. This, the intercostal sprout, represents the ramus anterior of the adult vessel. Its future great growth makes it the chief portion of the final vessel, but embryology shows plainly that the posterior ramus is the parent, and, again, that of the branches of this posterior ramus, the spinal branch is the primary or parent one and others (rami cutanei et museulares) secondary branches of it. From their origin to the point of division into posterior and anterior rami, then, the intercostal and lumbar arteries represent the original dorsal segmentals, but beyond the latter points they are entirely new and secondary formations. One may compare the above figures of the dorsal segmentals of embryos with the schema which I give in Fig. 438 to represent the adult.

Mall (1898) has shown that in the 16 mm. embryo anastomoses connect all the intercostal and lumbar arteries among themselves

638 as well as with the subclavian above and the femoral below. In this way, then, arise the a. epigastrica inferior and the a. mammaria interna, and along with the rectus, nerves, and ribs shift later into the mid-ventral line (Fig. 441). He thus explains the formation of the superior intercostal artery: "The descent of the heart into the thorax on the inside with the descent of the arm over the clavicle on the outside of the body causes great tension on the upper intercostal arteries, and favors the new formation of blood-vessels in

Fig. 441. — Arteries of the trunk in a human embryo 16 mm. long, showing the formation of the internal mammary and deep epigastric arteries. (Mall collection, 43.) (After Mall, Johns Hopkins Hospital Bulletin", 1898.) a more direct line. This is the reason why the main branch of the superior intercostal is a secondary and direct artery from the subclavian." Whereas the first two intercostals passed dorsal to the sympathetic chain originally, they now pass ventral to it.

Concerning the development of the muscular rami which belong to the dorsal segmentals little is known.

The cutaneous rami, though at one time thought to develop equally and symmetrically (Manchot, 1889), do not do so, as Grosser (1905) has recently been able to show. In fact, the segmental symmetry of these vessels is quite completely destroyed in the adult.

It is entirely probable tbat in tbe early stages of development tbe twigs which represent the blood supply of the skin are arranged perfectly symmetrically and

DEVELOPMENT OF THE VASCULAR SYSTEM. 639 segrnentally. They doubtless correspond accurately with the segmental cutaneous nerve branches. Both, passing out from their source, find their territory of distribution opposite them and at the same level. But the skin does not keep its relation with the skeleton, but shifts over it, dragging, as it were, its nerves and vessels with it. Thus it happens that hi the adult the segmental vessels and nerves no longer supply the skin area opposite them. Since in the thoracic region this shifting is chiefly caudalward, the cutaneous nerves all supply territories lying below their points of emergence from the intervertebral foramina. The arteries, however, though tending to follow the same law, also acquire new connections with the skin territories secondarily opposite them, and accordingly also supply besides their own proper segmental area territory which originally belonged to the adjoining more cranial segments. Such a departure probably does not obtain in the nervous system, where we may perhaps rely on the innervation of a skin territory to reveal its primary segmental position. In the case of the vascular system the departure is doubtless due to the tendency of a blood current to take the shortest possible path — a fundamental law in the development of the vessels. Some others accomplish this shorter path by the employment of anastomoses normally existing between the various cutaneous rami, and so come to course not only downward with the nerve of their own original segment, but also directly outward with the cutaneous nerves of contiguous upper segments and emerge with the latter into the skin. The original segmental skin arteries of these more cranial segments thus vicariously supplied may no longer play any role in the supply of the skin and in this way the number of actual skin vessels is reduced. Another cause, besides this shifting and secondary assumption of a shorter path, operates to disturb a primary segmental symmetry in the skin vessels. This also is fundamental in the development of the vascular system — the tendency of favored vascular channels to annex contiguous ones. Such a tendency is shown to a remarkable degree in cases of certain twin embryos, where we appear to have a contest between the two hearts. In the skin plexus the favored channels supplying this net enlarge at the expense of others, and this may result in the complete assumption of the territories of some three original skin rami by the vessel originally belonging to only one. It is probable that this tendency would operate in the absence of any shifting of the skin even though it is encouraged by the latter, for it is unlikely that exactly equal conditions should obtain in the case of supply of all the segmental skin areas, and a disproportion once established is rapidly exaggerated. This is without doubt the reason why both the posterior rami (it. cutanei dorsales mediales et laterales) of a particular vessel seldom persist, usually the medial rami alone persisting in the upper segments and the lateral rami in the lower ones.

The further history of the anterior spinal artery may be briefly given here. 40 His (1886) had noticed that in the human embryo of 18 mm. the single anterior spinal artery of the adult was finally present, and indicated that its definitive singleness was attained by a medial dislocation and fusion of the two primitive trunks, a process typified, for instance, by the well-known fusion of the two aortse. This view has never rested on any embryological evidence, Kadyi (1889), Hoffmann (1900), and others merely accepting it tentatively, following His. Although such a fusion seems to be actually the case in the elasmobranchs (Sterzi, 1904), in the higher vertebrates, and especially in all the mammalia, a 40 Sterzi has pointed out that the condition of paired anterior spinal arteries or a " tractus arteriosi primitivi," is never developed in the mammals to the degree seen in the birds. In the latter class they form large, much stronger and less transient trunks, — e.g., lasting from the third to the twelfth day in the chick. It is interesting to note that this condition is definitive in the cyclostomes.


series of more elaborate changes must occur before the single vessel is formed. These changes do not involve a fusion process, but consist essentially in the selection of one of the possible paths offered by the primitive vessels and a plexus which has sprung up between them. The single definitive vessel may thus be unilateral, median, or even oblique in origin (Sterzi, 1904, Evans, 1909). In the first case the adult vessel represents one of the original primitive paired vessels, in the other cases it is formed from the median plexus which connects the two primitive vessels." 8 The single anterior spinal begins to be formed in human embryos when a length of about 15 to 16 mm. is attained. The irregular, " vacuolated " character of the young primitive trunk (Fig. 440, E) betrays its origin from the original plexus, as elsewhere in the developing vascular system.

Variations. — The studies of Kadyi (1889), Burrows, 41b and others show that the form of the adult anterior spinal artery often bears the stamp of its method of origin, being median in some areas but in very many others truly right or left sided. In some areas it even retains its original plexus character (circuli arteriosi medullares), and in others consists of two strong parallel trunks which again unite, — e.g., Kadyi (1889), Taf. 3, Fig. 11.

His stated that the double aa. sulci were later shifted together in the midline, but this does not rest on evidence differing from that for his statement of the fusion of the anterior spinals. Usually, indeed, the aa. sulci or centrales are distinctly separate in man, even in the adult (Kadyi), thus disclosing their original paired origin from the primitive anterior spinals : a thing which Kadyi first discovered in man, Hoche (1899) in the rabbit and dog, and Sterzi has recently shown from many other instances to be the general mammalian plan.

Even in those rare instances in which some of the aa. centrales have a common trunk, this does not arise from fusion of the two original ones, but from the development of an anastomosis between these and the persistence of only one of the two penetrating trunks below the level of the anastomosis, as is normally the case in the birds (Sterzi). (Vide Sterzi's figure, page 311.) The aa. centrales are evident in chick embryos of the 96th honr and in sheep embryos of about 6 mm. In human embryos of 10-11 mm. they form two distinct rows of delicate vessels which enter the cord at the margin of the primitive ventral sulcus and, anastomosing on each side among themselves, produce two vertical or dorso-ventral planes of capillaries. These two rigid planes of capillaries form a striking picture of the internal circulation of the cord at this time.

4 ' a Sterzi was the first to show that the anterior spinal artery usually seen in the adult is only formed after the appearance of a series of anastomoses between the two parallel primitive trunks. The final vessel, according to him, may in some regions be derived from the left primitive vessel and in other regions from the right one, according to chance. The development of the anastomoses between the two primitive vessels permits the branches of that one destined to perish to be taken over by its more successful neighbor. Probably the usual anterior spinal is thus really unilateral in origin. At the same time, however, another plan may be followed in some areas. The anastomoses between the two primitive anterior spinals may become so large and numerous as to completely destroy in places the paired character of the arterial channels of the ventral cord surface and in such areas the cord is nourished by a rather wide median arterial plexus, from which later an exactly median vessel can emerge (Evans).

4 ' b Burrows, M. T.. unpublished observations.

DEVELOPMENT OF THE VASCULAR SYSTEM. 641 This is the earliest method of blood supply of the cord in all the higher vertebrates, a sole exception being made for the urodelous amphibia, in which the first cord vessels penetrate from the lateral surfaces (Sterzi).

The further development of the cord vessels is as follows: Some time after the entrance of the aa. centrales into the cord, other vessels also penetrate it from the lower lateral surfaces opposite the level of the dorsal margins of the anlagen of the ventral gray columns (aa. periphericce). For a while, although both these ventral and lateral penetrating vessels exist, the dorsal two-thirds of the spinal marrow is still non-vascular. The whole lateral sides of the cord and its ganglia are quickly covered with the capillary plexus, but few if any sprouts have ventured on to the dorsal surface ( 7 mm. pig embryos) . Thus the cord presents the remarkable condition of a close capillary investment everywhere save on its upper surface, which is as yet non-vascular. However, this surface is now rapidly covered, at first by delicate transerse capillaries which bridge the gap just as they do at first between the primitive anterior spinals. Gradually then a close mesh is formed here. The gray matter of the cord is better and better supplied by secondarily arising penetrating arteries, which may arise as far dorsally as just beneath the posterior nerve roots (sheep embryos of 10^ mm.). Eventually the aa. periphericae exceed in importance the original aa. sulci, an event which occurs not only in man, but also in the rodents, artiodactyls, perisodactyls, and carnivores, in all of which the peripheral penetrating arteries come ultimately to supply the greater part of the cord substance. In the chiroptera and insectivores, on the other hand, the original ventral segmentals remain always the : chief arterial supply of the cord. The white matter of the cord is always supplied late, it remaining practically non-vascular in sheep embryos until a body length of almost 50 mm. is reached. Gradually there develop on each lateral half of the cord four longitudinal anastomotic chains; the first to arise and more important of these forms at or just medial to the line of exit of the posterior roots (sheep, 50 mm.). This is the posterior spinal artery of descriptive anatomy (tractus arteriosus postero-lateralis of Kadyi), and corresponds to the tractus arteriosus lateralis of most mammals. Next, a similar but weaker anastomosis develops along the line of exit of the ventral nerve-roots (tractus arter. ventro-lateralis) (tractus arteriosus antero-lateralis, Kadyi). Finally, anastomotic arterial chains are established dorsal to the dorsal roots (tractus arteriosus posterior, Kadyi), and opposite the ligamenta denticula (tractus arteriosus lateralis), the latter being peculiar to man and the apes. Of the various longitudinal venous trunks which develop, the order of establishment is similar to that for the arteries, the ventral, lateral, and finally dorsal appearing successively.

Anomalies of the Dorsal Segmental Arteries. — As regards their manner of origin from the aorta, the dorsal segmental arteries show two main types of anomaly. They may (1) either disappear completely on one or both sides, their branches being taken over by the adjacent cranial or caudal segmentals, or they may (2) fuse with the vessel of the opposite side into a single median stem, a process normal to the ventral segmentals (vide infra).

Examples of the first type of anomaly are not infrequent in man, Krause having recorded cases in which as many as four interstitia intercostaliawere supplied by a single intercostal artery. It is interesting to note that such a condition occurs on one or both sides in the normal development of certain fish, amphibia, and birds. The second type of anomaly in which the two dorsal segmentals of one and the same segment fuse to a common stem is also common in man. Ernst has recorded a remarkable case hi which all the intercostal and lumbar arteries arose in this way, — i.e., for each segment from a single median trunk. Broman has found this second type of anomaly occurring in instances in the early embryo (13 mm.), and advances the notion that it occurs through an actual fusion rather Vol. II.— 41

642 than through the atrophy of one of the pair. 42 * Many years ago Krause emphasized that the two places in which this anomaly was commonest were in the lowest intercostal and lowest lumbar regions, and Broman suggests that this is connected with the fact that the aorta? first fuse in the lower thoracic region and that a marked fusion process, normally bringing the roots of the two common iliacs together, occurs in the lower lumbar region. Common stems are normally produced in the ease of some or all the dorsal segmental pairs in some mammals, — Lepus (Ernst), Halichoerus (Hepburn).

Arterial arches, cut off

Aorta deseendens dextra

7th dorsal branch

, _^. Arteria cceliaca primitiva

Arteria omphalomesenterica

Arterise umbilicales Arteria caudalis dextra

Cranial root of a. umbilicalis | Caudal root of a. umbilicalis 21st dorsal branch Fia. 442. — Reconstruction of the aorta and its branches in a human embryo 3.4 mm. long. (After Broman, 1908.) The Ventral, Segmental Arteries. (Gut Segmentals, Yolk Segmentals, "Visceral Circle" [Mackay]). — The first branches to be given off by the aortae, if we except the precocious and immense umbilical arteries, are those which course on to the primitive gut and the yolk-sac. Here the primitive aa. vitellinae were first seen in the human embryo by Mall (1897).

Bischoff (1842) has usually been given credit for the discovery of the row of yolk arteries given off by either aorta; his observations were made on the rabbit.

-° a But see Hochstetter, 1911.

DEVELOPMENT OF THE VASCULAR SYSTEM. 643 Von Baer (1827), however, had preceded him, for in his " de ovi ruammalium et hominis genesi epistola " (Fig. VII a) he shows some six or seven pairs of yolksac arteries in a young dog embryo.

When the two aortae have met and fused, opposite ventral arteries are quite accurately matched, as is always the case with the dorsal segmental arteries, so that from the now single aortic tube there go off at many places pairs of ventral or gut arteries which are also often accurately segmentally (i.e., intersegmental^) placed.

It should be mentioned, though, that, while this is the case for most of the aorta's length, in its most cranial portion the ventral branches have perished before the aortic fusion has taken place, so that a condition of paired ventral vessels from the single aorta does not ever come about in this region, — i.e., in the territory of the occipital and six upper cerical segments.

The most cranial lying ventral branches are very transitory, and the very first of them have entirely escaped notice until recently. In the Mall embryo No. 391 (Dandy, 1910) possessing seven somites, the ventral or gut branches extend as far forward as the first intersegmental cleft (Fig. 408). By the time the embryo possesses fourteen somites (2.1 mm., Mall, 1897, Pfannenstiel III, N.T. 6) the most cranial ventral branches appear in the region of the fourth and fifth somites. In the embryo with twenty-three somites (Robert Meyer, No. 300, N.T. 7) the ventral vessels opposite the next three caudally lying somites are also in degeneration, so that the vessels near the beginning of the eighth somites constitute the first of the functioning series.

In the Broman embryo of 3 mm. (N.T. 11) (Fig. 422) the ventral vessels opposite the 7th cervical dorsal pair constitute the most cephalic of the series, and this pair is probably the most cranial of the ventral branches to persist long enough for fusion of the aortae to occur in their neighborhood. 42 By the time the embryo attains a length of five millimetres, all of these ventral pairs have given place to single median stems (Fig. 443). Broman (1908) believes this to take place first in the middle of the unpaired aorta and to have proceeded cranially and caudally from this point. In a human embryo of five millimetres which Tandler (1903) has described, all of the ventral pairs have "fused" and there exists a complete series of unpaired or median ventral segmentals from the seventh cervical to the second lumbar segments inclusive. Broman (1908) describes these vessels as representing in each case a fusion of the original segmental pairs, and not, as has been supposed (Thane, 1892, and others), persisting right or left members of the original pairs ; but it is possible, as

42 Whether the oesophageal arteries which Broman and I have seen in quite young embryos are remains of these cephalically lying original vitelline vessels or entirely new sprouts does not permit of determination.

644 Felix remarks from his study of the embryo of 23 somites, that this is often not the case, since here occasionally right members of the ventral pairs were already larger. The question is an open one.

Broman has attempted to explain the normal fusion of the ventral segmentals, in contrast to the persistence of the paired condition which the dorsal segmentals

Aorta descendens dextra

Aortic arch

Truncus arteriosus

n s "E « t r. o T a o a o m

terica ._ superior

A. umbilicalis dextra


A. cauda Caudal root Cranial root Fusion gaps 23d right dorsal branch

Fig. 443. — Reconstruction model of the aorta and its branches in a human embryo 5 mm. long. (After Broman, 1908.) The cranial end of the right mesonephros and the position of the metanephric anlage are indicated by dotted lines.

exhibit, by affirming that the ventral vessels are from the very beginning placed nearer each other than are the two dorsal stems. This statement, of course, will not hold, as can be seen from the study of younger embryos than were at his disposal (Fig. 444). The coalescence of the ventral segmentals is doubtless connected with those forces which pull the intestine farther away from the aortic wall to produce the dorsal mesentery.



It is quite possible that the seventh pair of ventral segmentals remain longer than those above them just because they function as one of the roots of the cceliac artery. At the stage of five millimetres, although the series of mid-ventral segmentals may be uninterrupted, some of the members of the series are already much exaggerated over the remainder and enable us to recognize them as forming the cceliac and omphalomesenteric arteries respectively (Fig. 445). The former vessel arises by two roots from the seventh and eighth ventral segmentals and, coursing ventrally, forks, the two branches being traceable forward toward the portion of the alimentary canal from which later the stomach and liver are respectively derived. The omphalomesenteric artery is by far the largest of the ventral series, and, while its main trunk is the con

Caudal end of the 3d somite


Centre of the 3d somite


4th ventral segmental artery

V. urnbilicalis

Yolk-sac opening of the' gut



Fig. 444. — Cross section of a human embryo of 7 somites, showing the primitive ventral (segmental) branches of the aorta. The yolk-sac is so spread out that these branches appear as lateral derivatives of the aorta, although later ventral. (After a drawing kindly placed at my disposal by Dr. Walter E. Dandy.) tinuation of the thirteenth segmental vessel, the four ventral segmentals cranial to this also share in giving origin to it, for they are connected with this artery by a series of longitudinal anastomoses. As can be seen from Fig. 445, the omphalomesenteric artery splits on reaching the intestine and surrounds the latter at its junction with the ductus omphalo-entericus, with an arterial ring, before proceeding on its way to its final field of distribution on the yolk-sac. Fig. 446 shows conclusively that the left limb of this ring has atrophied, since the artery now passes entirely on the right side of the gut.

Anomalies. — Sometimes a considerable part of the old omphalomesenteric artery persists in those rare cases of the most primitive type of Meckel's diverticulum. In such cases what is undoubtedly the original artery courses beyond the gait and it? diverticulum to the umbilicus, and a dotprmination of nu which side of the

646 gut the vessel courses will disclose whether the right or left limb of the early arterial ring has persisted. All of the more advanced types of the diverticulum, in which the process is merely supplied by an unusually strong vessel but in which the old trunk cannot be identified with certainty, must be inadmissible for the determination of this point, for the diverticulum is a healthy functioning pocket of the bowel and as such could have secondarily attracted for its supply branches from the vessels of either contiguous wall of the intestine. 43

Ductus omphalo-entericus

A. omphalomesenterica

Ductus Wolffii

7th dorsal segmental artery

A. umbilicalis

Fig. 445. — Sagittal reconstruction showing the aorta and its branches in a human embryo of 5 mm. (After Tandler, Anat. Hefte, Bd. 23, p. 192, Fig. 1.) • Opposite the lower colon, no one of the ventral segmental arteries is especially enlarged above its fellows, and the equal part which all of them play in the nourishment of this part of the bowel

    • It is of interest to note that Allen (1883) some years ago pointed out that remnants of both the a. and v. omphalomesenterica are normally found in the newborn of the cat, dog, and guinea-pig in a strand of tissue which reached the navel.



prevents us from identifying any one of them as the a. mesenterica inferior. Nevertheless, in an 8 mm. embryo the latter artery is apparent as the 20th ventral segmental (Broman, 1907).

In the succeeding stages in the life of the embryo, the vessels which we must recognize as the cceliac, superior mesenteric, and inferior mesenteric respectively are all found at successively lower levels on the aortic wall, a fact which is to be correlated with the descent of the intestinal viscera (their territories of distribution) into the abdomen. This highly interesting phenomenon, the so

9th dorsal segmental artery

A. coal.

A. omph.-mes.

A. mes

Fig. 446.

-Sagittal reconstruction showing the aorta and its branches in a human embryo measuring 9 mm. (After Tandler, Anat. Hefte, Bd. 23, p. 197, Fig. 2.)

called "caudal wandering" of the visceral arteries, was first discovered by Mall (1891), and has since been abundantly confirmed and extended by the studies of Tandler (1903) and Broman (1908). The subjoined table shows the position of these vessels in a number of human embryos during the time of their migration (p. 648).

The cceliac artery thus wanders from the seventh cervical to the twelfth thoracic segments, a displacement of some eleven segments, and the superior mesenteric artery almost equally as far (ten segments, second thoracic to first lumbar) ; whereas the inferior mesenteric artery wanders through but three segments (twelfth thoracic to third lumbar). The great change which the levels of origin of the first two vessels undergo, in contrast to the slight one of the third, is readily intelligible from the proportion

648 ately great dislocation which the upper part of the alimentary tract undergoes. All of these vessels usually attain their adult levels by the time the embryo is 17 mm. long.

This shifting of the intestinal arteries is not produced by a displacement of the aorta on the vertebral column, but is an actual

Length of embryo. Position of a. coeliaca. ' Position of a. mes. sup. Position of a. mes. inf.

1 4.9 mm. . .

2 4.5 mm. . .

3 5 mm ....

4 o mm ....

5 6.75 mm. .

6 7 mm. . . .

7 8 mm ....

8 9 mm 9 9 mm . . . .

10 10 mm. . .

11 10.3 mm. .

12 11 mm. . .

13 11.7 mm. .

14 11.7 mm. .

15 12.5 mm. .

16 13.2 mm. .

17 14 mm . . .

18 14.5 mm. .

19 14 mm . . .

20 14 mm. . .

21 15.5 mm. .

22 16 mm . . .

23 16.2 mm. .

24 16 mm . . .

25 16 mm . . .

26 17 mm . . .

27 19 mm . . .

28 19 mm. . .

C. 7 Betw. C. 8 and T. 11 C. 7andC. 8 C. 8 and T. 1 T. 2 T. 5 T. 2 T. 4 T. 4 T. 8 Betw. T. 7 and T.8 T. 6, 7, 8 Betw. T. 7 and 8..

T.9 T.8 T.8, 9 T. 10 T.9, 10 T. 10 T.ll T.ll T. 12 T. 11 T.ll (lower part). T. 12 (upper part) T. 12 T. 12 T. 12 (lower part) .

T. 1,2,3,4.

2 and T. 3... 1,2,3,4,5..

4 and 5 5 and 7

Betw. T. 5 and 7.

T. 4, 5, 6 T. 5, 6, 7 T. 6, 7 T.9, 10 T.9, 10 T.8, 9 T.9 T. 10 T. 10 T. 10, 11 T. 10, 11 T. 11 T. 11 T. 12 T. 12 T. 12 T. 12

T. 12 (upper part) T. 12 (lower part) .

L. 1 L. 1 L. 1

L. 1 T. 12 T. 12 L. 1, 2 L.2 Betw. L. 1 and 2 L.3 Betw. L. 1 and 2 L.2 L.2 L.2 L.2 L.2 Betw. L. 1 and 2 L.2 L.2 L.3 Betw. L. 2 and 3 L.2 L. 2 (lower part) L.3 L. 3 L.3


Ingalls. 44 Broman. Tandler. Broman. Keibel and Elze. Elze. Broman. Tandler. Tandler. Broman. Broman. Broman. Broman. Broman. Tandler. Broman. Broman. Tandler. Author. Tandler. Author. Broman. Broman. Author. Author. Tandler. Broman. Author

shifting of these ventral branches when compared with the dorsal branches of the same trunk. 45

44 " Zwischen dem f iinf ten und seehsten Rumfganglion findet sich ein bis an den Darm verfolgbares Gefass, das vielleieht als a. mes. inf. anzusehen ist." (Ingalls.) 43 The exact manner in which this wandering of the gastro-intestinal vessels is accomplished has not as yet been established. Undoubtedly one possible method in early stages is by means of the anastomoses which connect the ventral vessels. This, however, will only account for very early shiftings, for the studies hitherto made show that very soon there may not be a single other vessel between the points of origin of the three chief vessels (e.g., Tandler's embryo K.S.). Consequently other methods have been called on to explain this caudal wandering. These are — 1. That it takes place through the formation of special non-segmental anastomoses between the wandering arteries and the aortic wall below them, with the ensuing atrophy of the older roots. The chief evidence in favor of this view consists in the frequent presence of non-segmental roots of origin for these vessels. The original roots being all supposedly segmental, any non-segmental position for the vessel is explained by the acquirement of secondary non-segmental roots. Such a view overlooks the fact that even in the beginning non-segmental ventral branches are present (see, for instance, the vessels in Broman's Fig. 1, page 646).

DEVELOPMENT OF THE VASCULAR SYSTEM. 649 Regarding the development of the peripheral branches of these arteries in man almost nothing is as yet known. 46 Tandler has identified the a. pancreatico-duodenalis superior in an embryo 13 mm. long (N.T. 57). At 15.5 mm. (Mall's collection, 390) the coeliac axis possesses the following branches: a. phrenica inferior, a. gastrica sinistra with oesophageal rami, a. hepatica with its a. cystica (strongly developed), a. pancreatico-duodenalis superior, and a. lienalis (Fig. 447).

Interest attaches to the development of the ventral branches which the adult aorta is known to send to the (esophagus, especially as to whether these also are descended from the early segmental branches. Some of these aa. oesophageales have moreover been identified in relatively early stages, but they are apparently new formations. 47

2. That it takes place through an active ventral wandering, by which it is understood that the caudal wall at its junction with the aorta bulges itself out, while the cranial wall at a corresponding place is taken up by the aortic wall. There is no evidence for this view.

In discussing the subject it is to be pointed out that the cceliac and superior mesenteric arteries have their roots in an uninterrupted chain of anastomosing vessels, and there is no a priori reason why the vessel functioning as the superior mesenteric in one stage may not subsequently be used as the coeliac channel. As the area of distribution of one of these vessels shifted caudally, the blood stream could adapt itself to a more direct path by the employment of these anastomoses which enable it to come from successively lower segments of the aortic wall.

It seems to me most probable, however, that the identity of the three main vessels is established permanently very early, and that the great shifting is due to an entirely different -phnomenon, — namely, to the unequal growth of dorsal and ventral walls of the aorta. Attention may be called here to the remarkable shifting undergone by the fourth aortic arch, for instance, compared with the dorsal segmental vessels, and yet the arches have not been thought to climb down by special secondary roots, etc.

48 Fransen has studied the branches of the a. mesenterica inferior in two human fetuses between the eighth and ninth month. The six chief branches which he finds going off from this artery he interprets not as the usual branches of the third order, but as original ventral segmentals from the aorta and sacralis media, which subsequently became united through a longitudinal anastomosis (the ascending and descending rami of the a. mesenterica inferior), whereas the root portions die. There is nothing embryologically to establish this claim. These lower ventral segmentals do not exist long enough to leave a permanent trace in the mesenteric plexus. Like the earliest limb vessels they are usually of a capillary nature. The establishment of the inferior mesenteric artery rearranges the whole vascular pattern of its territory of distribution, and the six branches to which Fransen refers came out of this plexus.

" In an embryo of 4.5 mm. Broman has identified three of these vessels risin? from the right aorta opposite the third and fourth dorsal segmental vessels, and reports them in embryos of 10.3 and 14 mm. I have myself seen them in embryos of 10 and 14 mm. In the embryo of 15.5 mm. shown in Fig. 447 they are shown as delicate twigs opposite the sixth and seventh thoracic segments, and have been seen in this location or slightly lower in five other embryos measuring from 19 to 23 mm. (Numbers 229/368, 108, 57, and 382. Mall's collection). In the older of these embryos they were represented by a fairly strong vessel opposite the eighth thoracic segment.

650 A. comm._ post.

R. supraorbi talis (arterise stapediales) A. ophthalmica

A. infraorbitalis

-A. basilaris] A. spinalis £»

A. stapedial A. occipita

A. maxillari A. lingualis A. thyr. bui A.temp. suy A. masdllar?

\ — ! — A. vertebralie A. thyreocervicalis

Rami o3BOphageales a. gastric, sin A. gastrica sin A. cystica A. pancreaticoduoden. sup.

A. epigrastica inf .• A. femoralis A. pudenda interna —

j. L..A. pulmonalii

-4th thoracic, segmental ar

- -Aa. cesophagj A. phrenica inf.

A. coeliaca — A. mesenteries su A. lienalis A. mesenterica inf. -3d lumbar segmental ai A. sacralis media -A. ischiadica

a series of sagittal sections. (No. 390, Mall collection.)

DEVELOPMENT OF THE VASCULAR SYSTEM. 651 As far as I know, nothing has been ascertained concerning the development of the bronchial arteries. In the embryo of 15.5 mm. (Fig. 447) three ventral branches of the aorta are seen to constitute aortic vasa vasorum.

The main branches of the mesenteric arteries are formed very early and can be identified in mammalian embryos well under 10 mm. in length. From the time of the earliest existence of the ventral segmentals, the gut is supplied with capillaries, and in the early embryo these form a close plexus in the tissues of the simple intestinal tube.

The earliest capillaries plexify in a fairly definite plane which corresponds to the future submucosa. This tunic — the so-called " tunica vasculosa " of the older anatomists — contains, as is well known, the chief plexus of intestinal vessels in the adult; there the chief vessels of the intestinal wall are found, and it is from them chiefly that the muscular rami and all of the mucosal rami are derived. This fact finds a better comprehension from the history of the vascularization of the gut wall, for in the submucosa the earliest and hence oldest vessels are found. From this layer of vessels, with the progressive development of the muscularis and the mucosa, there sprout out the rami which nourish these tunics. When the first villi are formed they receive simple capillary loops and sprouts; from the capillary plexus of the older villi, the villous arteries and veins are formed. The increase in complexity of the proper intestinal vessels proceeds from above downward, just as does the development of the intestinal walls and especially the villi; the vessels of the small intestine much precede in complexity those of the large bowel, and the latter portion, for a long time smaller in girth, remains supplied only with a single, simple, submucosal net at a time when the small gut has manifold muscular and mucosal rami.

Anomalies. — The cceliac and superior mesenteric arteries sometimes arise from a common trunk — the so-called " cceliaco-mesenterica," Rathke. This is an entirely normal condition in the Anura, some of the Chelonia and Lacertilia, and some of the Mammalia (PhocaBna [Cuvier], Talpa [Tandler], Echidna [Hyrtl], etc.). The formation of such a trunk has been interpreted as due to the approach and fusion of the cceliac and superior mesenteric arteries (Howes, Klaatseh, Fransen, etc.). Tandler (1904), however, has studied the embryonic development of Talpa, in which this occurs as a part of normal development. He finds a strong longitudinal anastomosis between the various early segmentals of the cceliac and superior mesenteric group. Only one of these early segmentals remains as the permanent trunk, and it has as its chief cranial branch a longitudinally coursing vessel, which is doubtless the old longitudinal anastomosis between the segmental series, the cranial members of which have now degenerated. From this longitudinal vessel the gastric (sinistra), hepatic, and splenic arteries are later distinguished as arising. The main part of the permanent trunk is the omphalomesenteric channel; in this way, then, the anastomosis enables the latter vessel to take over the branches which usually belong to the cceliac. Tandler has applied these findings to explain also the anomalous occurrence of an a. cceliaco-mesenterica in man. If his schemata are interpreted liberally as signifying any mesenteric anastomoses by virtue of which one vessel can take over the whole or part of its neighbor, they deserve to stand as the most reasonable and plausible explanation for these anomalies. It is significant that it is always the stronger vessel — the a. mesenterica superior — and never the weaker cceliac which performs the annexation, a fact in conformity with our general ideas of the method of development of the vascular system. Tandler in fact recognizes a general anastomosis between the branches of aorta in this


region, constituting, as it were, a general cceliaco -mesenteric complex. Normally there occurs a later separation of the cceliac and mesenteric systems. Broman, on the other hand, thinks that from the earliest time at which they can be recognized these two vessels with their multiple roots are entirely separate; this is because the human material hitherto explored has not revealed a complete chain of anastomoses between the two vessels, as it has in Talpa. The limitations of method of attack here make it probable that these vessels can not always be seen and that future researches will show them present. If they are not present, another method of formation of a truncus cceliaco-mesentericus may be the correct one; this is the active outgrowth of a wandering root from the cceliac which attaches itself to the superior mesenteric rather than the aorta (Broman).

The rather commoner, longer anastomoses between the cceliac and upper mesenteric arteries are doubtless more secondary developments from the plexus in the primitive mesentery. (In this category are to be placed the cases reported by Aeby, Biihler, Fawcett, Tandler, Thane, Toldt, and others.) The superior mesenteric artery has also been reported as taking over the field of the inferior mesenteric (Fleischmann, 1815), but this is doubtless an anomaly of the greatest possible rarity, because the lower vessel is initially so far removed from the superior one as to be from the beginning a far more effective supply for the bowel which is opposite it.

Lateral Branches. (Nepkric Segmentals, Intermediate Arteries (Mackay), etc.). — Mention has already been made of the occurrence of primitive lateral branches of the aorta in human embryos of 15 and 23 somites (see Fig. 436). The relation of these vessels to the lateral branches of the aorta present in embryos of 4 to 5 mm., and which are now clearly concerned in the supply of the Wolffian body, is not clear, and will not be so until intermediate stages are possessed. I shall discuss here only the latter arteries, which we may designate simply as lateral branches of the aorta or the mesonephric arteries.

His (1880) first observed multiple branches of the aorta supplying the mesonephros in a seven millimetre embryo, and Mall (1891) emphasized the tendency of these to be segmentally arranged in early stages. 48 Broman has recently given a more extended account of them and their fate in a series of embryos, and I follow him.

At first, when the "Wolffian bodies are relatively small, the number of mesonephric vessels is correspondingly small and these 48 Tandler has confirmed this tendency for a segmental arrangement of the mesonephric arteries, but the studies of Broman, Ingalls, Elze, etc., show that many non-segmental arteries exist either from the beginning or as a result of shifting of original ones and we must admit that the metameric arrangement of the Wolffian body arteries is soon completely lost. Hochstetter has called attention to the fact that the mesonephric vessels in Selachians are described as coming from the segmental body wall arteries (Dohrn), and in Amphibia as being true segmental offshoots of the aorta (Semon). He is of the opinion that the corresponding vessels in amniotes were also segmentally arranged in correspondence with segmental mesonephric glomeruli, each of which had its own artery. Actual observations on amniote embryos which will support this have not yet been made.



come from only the middle portion of the aorta (2d to 8th thoracic segments) ; but when, at the end of the first month, the mesonephros reaches its greatest development, it receives many direct branches from the aorta at levels cranial as well as caudal to the original ones. The following table will indicate this :

Length of embryo.

Level of origin of mesonephric arteries.

  • No. of.

mesonephric arteries Observer, on each side.

5 mm 5 mm 7 mm 8 mm 2d to 8th th. segments 1st to 12th th. segm 8th cerv. to 12th th 8th cerv. to 12th th. segm 7 Broman (1908) .

13 Tandler (1903).

14 Mall (1891).

20 Broman (1908).

The last vessels added to the series appear to grow out from the level of the first lumbar to second lumbar segments in embryos of 10 millimetres. These indeed are destined to persist in the adult representatives of these arteries, for all the remainder atrophy by the time the embryo is from 16 to 19 mm. long. When the sex glands and the adrenal arise, they are supplied by branches from many of the neighboring mesonephric arteries.

Gradually the sex gland loses all but a single one of its many arteries, and this is the branch from the mesonephric vessel opposite the second lumbar segment. The atrophy of the Wolffian body permits the entire blood stream in this artery now to supply the sex gland, and thus the a. spermatica interna appears to be a direct branch of the aorta (Hochstetter for mammals, Broman for man). 49 The branches of the mesonephric arteries to the adrenal gland are originally many (6 at least), and come off from the higher members of the series, — e.g., in a 10 mm. embryo, from the mesonephric arteries arising from the sixth to eighth thoracic segments. But eventually, with the relative descent of the adrenal, it acquires branches from the Wolffian body arteries at lower and lower levels. At last in 16 mm. embryos the adrenal arteries are branches of three mesonephric vessels near the first and second lumbar segments. 50 With the atrophy of the Wolffian body, these three adrenal arteries persist and consequently take over the entire blood current, thus appearing as three independent branches of the aorta. Before the adult state is reached, the upper and lower members of the series of three adrenal vessels acquire important secondary con

49 Along with this goes the fact that the recognizable rudiments of the Wolffian body in the adult — the epididymis or epoophoron — are naturally supplied by the sex gland artery.

50 It is to be noted that at this stage these are at last the only mesonephric arteries existing, with the exception of the very last member of the series — that of the second lumbar segment — which sends a branch to the sex gland.


nections, for the latter comes to supply the permanent kidney (a. renalis), 51 and the former the diaphragm (a. phrenica inferior). These secondary fields for the upper and lower adrenal vessels soon exceed in importance their adrenal territory, and so, in the adult, we only speak of the upper and lower adrenal arteries as small branches of the large renal and inferior phrenic vessels, though embryologically the reverse is the case. The a. renalis soon takes a descending course, and only in the second half of fetal life does it appear transverse. 52 Luna (1908) has shown that the a. phrenica inferior does not surpass its adrenal branch in size until about the seventh embryonal month.

As a result of all observations hitherto made, it may be stated that the permanent kidney in mammalian embryos certainly does not receive any large and readily appreciable arterial supply until its definitive position is reached. Hochstetter has stated that the vv. renales also wait such a time for their development. These facts, however, can not be taken to mean that the renal anlage possesses no circulation during the important early period of its development. For it can be shown, even from ordinary histological sections, that the kidney during this time possesses many small vessels in its walls, and Broman (1907) has recently traced connections between these and the posterior cardinal veins, on the one hand, and with the efferent Wolffian body veins, on the other. 53 This is not the only source of blood for the early metanephros, for injections of mammalian embryos (pig) indicate that its capillaries receive arterial blood from the a. sacralis media (Fig. 448) and inferior mesenteric artery. (See Jeidell, 1911.) Variations. — Supernumerary renal arteries have been known for a long time (Macalister [18831 records a case of seven), but until the embryology is accurately known explanations for their occurrence will be highly speculative, as they have been in the past. From the time of Meckel onward, there have been observers willing to postulate a hypothetical " splitting " of the usual single renal artery to explain this! (e.g., Kolster, recently). However, other observers have stated their belief 81 Hochstetter declares the a. renalis of other mammals to be a direct secondary outgrowth of the aorta, and the same history was described by Hill for the pig. The subject is worthy of reinvestigation in very early injected embryos.

62 Broman explains this by a descending course of the a. suprarenalis inferior at the time the a. renalis supplants it. Formerly these vessels were transverse, but after the closure of the diaphragm he thinks the latter successfully prevents any upward extension of the adrenal, and that adrenal growth from now on consequently pushes down its lower pole together with the a. suprarenalis attached there.

63 From such a finding Broman concludes that the post-cardinal venous blood flows to the kidney and is drained out again into the vv. revehentes of the Wolffian body; this would furnish a renal-portal system for the early metanephros comparable with the renal-portal system so well known in the ease of the mesonephros, and in accord with similar observations made some years ago by Hochstetter on the metanephros of reptiles.



in the derivation of supernumerary aa. renales from Wolffian body vessels, and Bromans derivation of the normal a. renalis from this source makes this explanation of multiple renals the most plausible. The abnormal origin of the renal artery in common .with other trunks is of some interest, inasmuch as we can now explain a large number of these embryologically. The inferior phrenic, adrenal, and sexgland arteries being derivatives of the original mesonephric vessels, all combinations in the origin of the former vessels may be expected. Thus the origin of the a. spermatica interna from the a. renalis is not uncommon, the common origin of a. renalis and a. suprarenalis inferior is normal, and the other adrenal vessels may likewise come from the renal. It is now possible, in view of new observations on the earliest blood supply of the metanephros, that certain types of origin

aorta abdominalis - v. card. post.

v. segmentalis dorsalis a. segmentalis dorsalis

v. segmerrtalisjlorsalis a. segmentalis dorsalis

,a. sacralis media ,v. ischiadica a. segmentalis dorsalis corresponding v. segmentalis dorsalis

v. segmentalis dorsalis a. segmentalis dorsalis

Fig. 448. — Arteries to the permanent kidney in a pig embryo 14 mm. long, after an injection of the living embryo. The arteries in question are the small upwardly-directed branches which arise from the a. sacralis media and the lateral plexus formed by the a. sacralis media. The same plexus is seen to give rise to the aa. segmentales dorsales on each side.

of the renal artery from lower sources — e.g., from the a. mesenterica inferior or a. sacralis media — represent the retention of its first vascular connections when the gland was pelvic in position. There still remain, of course, many remarkable anomalies of all these arteries which indicate entirely secondary shiftings or connections, — e.g., the origin of the a. spermatica interna from certain lumbar arteries. Broman has emphasized that the mesonephric arteries may come off at variable points from the lateral aortic circumference, many of them, in fact, being ventrolateral derivatives. It is easy to understand how, in the latter cases, in further growth the mesonephric artery may come to be incorporated with a contiguous ventral branch of the aorta. The most common instance of this is afforded by the common origin of coeliac and inferior phrenic arteries.

End Bkanches of the Aoeta (Caudal, Lower Limb, and Pelvic Arteries). — In all vertebrates in which the hind limbs are im


portant, the aorta does not appear to go over insensibly into the a. caudalis, but is rather drained of most of its blood by the mighty iliac branches, which we have come to speak of, in addition to the caudal vessel, as the end branches of the aorta. The simplest arrangement of the aortic end branches is that seen in man, and involves merely a tripartite division" into the two common iliacs and the a. caudalis (a. sacralis media). 54 In many mammals, including man, later shifting makes the sacralis media appear as a dorsal derivative of the aorta and not as its direct continuation, — e.g., in the human adult it almost constantly arises cranialward from the "bifurcation place" of the aorta. 55 In the human embryo we have seen that the tremendous importance of an early placental circulation has "pushed forward" the development of the umbilical arteries so that they much precede of course the appearance of limb arteries.

Studies on early embryos show that the umbilical artery is relatively farther cranial in position than it subsequently comes to be, — i.e., it appears to wander caudally. We have seen that thp primitive umbilical arteries possess many roots of origin from the aorta which are in fact only the caudal members of the general vitelline series (aa. vitelline). 56 64 Hoehstetter has shown that in some mammals, although this plan is originally followed, there subsequently occurs a disappearance of the common iliac vessels, so that the external and internal iliacs arise separately from the aorta (cat). Hoehstetter in 1903 felt inclined to explain this as due to a splitting of the aa. iliaea communes, Broman (1908), by a fusion of the umbilicals down to the point of origin of the external iliacs. Very recently now, Hoehstetter (1911) has subjected the matter to a careful restudy, and comes to the conclusion that the cat's truncus hypogastricosacralis comes through a, wandering upward of the origin of the a. iliacae externas from the wall of the a. iliacae communes to that of the aorta ; a similar thing apparently occurs as regards the a. iliolumbales which wander from the external iliacs to the aorta.

85 Young has attempted to maintain that the umbilical arteries really represent the original continuations of the aorta? which have fused only down to the point of origin of these vessels. He goes over into a hypothetical and poorly founded realm in declaring that the aortae thus bend around into caudal arches comparable with the aortic arches. He explains the a. sacralis media as a secondary branch, being much impressed with its dorsal origin from the aorta at a point cranial to the iliacs rather than at the exact division place. Nevertheless in development the sacralis media goes off at the point of origin of the a. umbilicales, and in addition behaves like the aorta in its position, dorsal segmental branches, etc. Broman explains the definitive origin of this artery cranial to the division of the aorta into its iliacs by assuming that the last part of the aortic stem is formed by a secondary fusion of the aa. umbilicales for a short space at their proximal ends. No evidence exists for this view, and if relative growth differences cannot completely account for the apparent cranial shifting of the sacral artery, we must assume a true wandering to have taken place.

  • In this connection it is of interest that in some mammals the omphalomesenteric artery first appears to take origin from the umbilical by a stem which leaves the

DEVELOPMENT OF THE VASCULAR SYSTEM. 657 In very early stages this caudal migration of the umbilical artery is unquestionably brought about by the caudal growth of the aorta itself together with its intestinal branches, the whole forming a plexus with which the umbilical arteries are constantly in relation and by means of which the blood to them gradually flows in more and more caudally placed ventral branches. Thus, in the Kroemer-Pfannenstiel embryo of 6 somites (N.T\ 3), these vessels arise at about the level of the future seventh or eighth segment, — i.e., the fourth cervical somite. In embryos measuring less than 4 mm. the artery is almost at the level of the first lumbar vessels. It is probable that the single or at most double roots which the a. umbilicalis possesses at this stage are its final ones which belong to the original vitelline series. These roots, however, are themselves displaced or "wander" caudally, so that in embryos of 5 mm. they are found at or slightly below the level of the third lumbar vessels. 57 They do not, however, constitute the definitive roots of these arteries, for, as Hochstetter (1890) some years ago showed for rabbits, the umbilical arteries of mammals next gain a more laterally placed root of origin from the aorta by the development of an anastomosis with the posterior limb arteries, whose origin from the aorta now becomes the root trunk of the umbilical artery, the original ventral umbilical root now atrophying. Hochstetter showed clearly that both ventral and lateral roots for the umbilical may exist for a short time coincidently {e.g., in rabbits of eleven days, two hours), and so form an arterial ring enclosing the Wolffian duct and ccelomic cavity, lateral to which the secondary roots and medial to which the primary roots course. Such a condition can be seen in human embryos of about 5 mm. (N.T. 16) as Keibel and Elze (1908) first reported and as may be seen from latter vessel and courses cranially toward the place where later strong arterial connections with the aorta enable the proper omphalomesenteric artery to displace it. Those are the conditions seen by Ravn (1894) in the rat and mouse, and I can report an almost similar phenomenon in early embryos of the pig. Here injections show that there exists for a time (7.5 to 9 mm.) a strong arterial route for the omphalomesenteric artery which arises from the a. umbilicalis and courses cranially to join the former vessel. These phenomena were quite unintelligible before we were aware, as we now are, that the entire vitelline-umbilical complex of vessels is originally one and the same system.

67 Broman has explained this caudal " migration " of the umbilical arteries by the successive development of " wandering " roots by virtue of which the artery acquires lower and lower connections with the aorta. As evidence of this he points to the double-rooted condition in which the artery may be found. This coincides with his explanation for the descent of the gut arteries. It is to be pointed out that many embryos do not show these multiple roots, and the appearance, even when found, is possibly merely an instance of inselbildungen. Disproportionate growth of the two aortic walls may again be responsible for this dislocation, or we may have to do with an actual active caudal migration of an individual trunk.

Vol. II.— 42

658 Felix 's Fig. 449, drawn from the Keibel embryo. In the embryo of 7 mm. (N.T. 28) only the secondary root is found, and the vessels are at their permanent location (at or slightly below the level of the fourth lumbar dorsal segmentals). 58 In embryos of this age, then, the strong umbilical arteries are found giving off, shortly after their origin from the aorta, a distinct branch, which courses somewhat downward and outward to the posterior limb, where it goes over into a capillary plexus. This is the primary artery of the limb, the a. ischiadica, and, while originally reaching the limb

Median root of the a. umbilicalis

a. umbilicalis

Lateral root of the a. umbilicalis

Renal bud

Fig. 449. — Reconstruction of the a. umbilicalis in a human embryo 5.3 mm. long. (Collection of Professor Keibel, No. 1420.) The umbilical arterj' is seen to arise from the aorta by three roots, a visceral and two parietal. (After Felix, 1910.) tissue without piercing the lumbar-sacral nerve plexus, at length the ventral growth of the latter makes this necessary in embryos of 9V 2 mm. (Elze, 1907).

Soon there also arises, from the upper side of the umbilical artery, the second vessel to the limb, a. femoralis, and we may now designate the umbilical trunk from the aorta to the femoral branch, the common iliac, for what is at first merely a femoral soon gives off the a. epigastrica inferior and other branches and consequently

68 According 1 to the investigations of Hochstetter (1911), they have, indeed, wandered to a position farther caudalward than that in which they are normally found in the adult, for in embryos of this age (6.5 to 10 mm.) the fifth lumbar arteries still usually arise from the division place of the aorta, whereas in later embryos, as in the adult, the aortic bifurcation is " pushed up " to lie opposite the fourth lumbar vessels, so that the aa. lumbales V can no longer be found coming from the aorta directly but are given off by the a, sacralis media. The latter vessel appears to wander cranialward independently, and so comes to arise from the dorsal wall of the aorta rather than at its exact division place; it may even be found giving rise to the aa. lumbales IV.

DEVELOPMENT OF THE VASCULAR SYSTEM. 659 comes to be the a. iliaca externa of the adult. The remainder of the umbilical trunk together with its a. ischiadica constitutes the definitive a. hypogastrica. Now the ischiadica is soon not merely that vessel, for in the 15.5 mm. embryo it gives off a prominent a. pudenda interna (Fig. 447). The root portion of the ischiadica from umbilical to this division place is consequently probably the great anterior division of the a. hypogastrica in the adult, and after the origin of the internal pudic comes to be the a. glutea inferior before at last the a. comes nervi ischiadici is reached. We are still without any series of observations on the development of the pelvic vessels.


For no portion of the vascular system do we stand in such need of a series of well-verified observations as we do in the case of the embryology of the extremity vessels. This field is of profound interest, too, from two stand-points : first, because the developmental history ought to furnish us with a key for the explanation of the manv anomalies which the limb vessels show and which have formed the basis for classic studies on the variation of the vascular system {e.g., Baader 1866, Ruge 1883, etc.) ; and, secondly, because enough has already been learned to indicate that the first arterial tree in the limb recapitulates in a striking way the simpler conditions which are definitive for some of the lower vertebrates (Zuckerkandl, 1894). The subject gains added interest also from another aspect, for from a closer study of the extremity vessels, Miiller (1903) and De Vriese (1902) in recent years have been led to advocate the idea of a capillary plexus ancestry for vascular trunks, in contrast to notions which had previously prevailed. Subsequent research has confirmed this general idea, extending it in some places and restricting it in others, as has already been mentioned. However our exact knowledge of the development of the subclavian tree is still scanty, and there is an even greater dearth of observations in the case of the lower limb.

Arm. — The earliest channels of an arterial source into the anterior limb buds are doubtless capillaries which arise directly from the lateral aortic wall at many points and anastomose profusely in the early limb tissue.

This stage has yet to he described for man, but may be shown clearly by injections of embryos of the chick and duck (Fig. 391). That it also obtains in the mammalia has been recently indicated by the reconstructions made by Goppert (1908) for the early subclavians in white mice (Fig. 450). Thus, as many as eleven of these earliest subclavians have been seen in the birds and five in the mammalia (Goppert). The capillary plexus which is formed by the anastomoses and further extension of these delicate vessels into the tissue of the limb is uniformly distributed in the blastema of the latter, save in a definite marginal

660 zone which constitutes a narrow non-vascular shell of mesenchyma lying beneath the ectoderm. The plexus is drained into the posterior cardinal and umbilical veins through a series of small venules, and later, after the survival of only a single subclavian artery, the well-known marginal vein of Hoehstetter is established.

Very soon after the outgrowth of these early multiple subclavians, changes occur which involve a disappearance of those vessels not arising at intersegmental points, so that the arrangement retained consists of two or more subclavian arteries which are located exactly opposite the dorsal segmental vessels in this neighborhood and are hence "segmental subclavians. " In the mammalia, including man, one of these segmental subclavians is always opposite the seventh cervical dorsal segmental vessel (according to Hoehstetter 's method of counting, the sixth), and others

V. card. C.

Fig. 450. — Reconstructions of the arterial system in the arm buds of embryos of the white mouse, S days old. (After Goppert, Verb., d. anat. Gessell., Vers. 22, Anat. Anz., 1908, p. 94, Figs, la and lb.) may have persisted at segmental points above or below this. 59 Very soon after the establishment of strictly segmental subclavians, these vessels are incorporated in common stems of origin with the dorsal segmental vessels, so that they no longer appear as direct lateral derivatives of the aorta, as was the case earlier, but become strong side branches of the dorsal segmentals.

All the stages just mentioned, however, are passed over by the time the human embryo attains a length of five millimetres, for at this stage only a single member of the early subclavian series remains to constitute the definitive subclavian artery, the vessel of 59 Thus, three segmental subclavians have been seen in the rabbit and mouse, and several cases of two segmental subclavians reported for man. The human ^-ens 16 and 17 of the N. T. possess subclavians from the sixth and seventh — «» 148 in Mall's collection has segmental subclavians from the irst thoracic segments. These observations show the possibility ubclavians in man. — i.e., from the last three cervical and first

DEVELOPMENT OF THE VASCULAR SYSTEM. 661 the seventh segment. It forms now the sole supply of the capillary plexus formerly nourished by multiple vessels and, after a short course to the root of the extremity, is soon resolved into a "spray" of many capillaries. As Miiller (1903) has shown, the main stem of the artery at this stage, while tending to be a fairly strong trunk, centrally located, often shows island-formations in its course, and eventually, before the true capillaries arise, becomes quite plexiform in character (Fig. 451). G0

Arterial island .,---'" \ formation \ i I I


Fig. 451. — Arm bud of a human embryo 5 mm. long, showing central arterial net. (After Miiller, Anat.

Hefte, Bd. 22, Taf. 25-26, Fig. 1.)

In the next stage which has been described, that of a 7 mm. embryo in the excellent study by Elze, but little change has occurred. No inselbildungen happen to occur into the proximal course of the artery, nor, apparently, is there any plexiform condition of the artery before the capillaries are given off.

In an embryo of 8 mm. Miiller found the subclavian giving off a branch just before the ventral nerve mass was penetrated; this branch continued for a short distance still medial to the ventral nerve, eventually plunging obliquely through the latter and joining the main stem, which has kept along the lateral side of the nerve; from this arterial loop, the main stem continues along the lateral

00 Within the tissues of the limb, then, the central arterial channel is not everywhere in the form of a single tube, but is rather constituted by an axial arterial plexus from which the capillaries are given off. The same character for this central nourishing channel of the early limb can be demonstrated by injections of the vessels in other mammalian embryos, and may be correctly taken to indicate that for a time the arterial current employs several instead of a single channel out of many available channels open to it by reason of the pre-existing general capillary mesh.

662 side of the nerve, and other fine vessels are given off to course just ventral to the dorsal nerve mass of the limb, as well as ventral to the ventral nerve.

This is evidently the condition occurring still in the 9.5 mm. embryo which Elze (1907) has carefully reconstructed, although the branch of the subclavian given off to continue medially along the ventral nerve does not anastomose with the main vessel, which, as in the previous stage, continues along the lateral side of the ventral mass, especially along the m. medianus; a more dorsally directed branch can be followed along the radial nerve (Figs. 452 and 453).

Arterial branch which pierces the plexus brachialis between n. cerv. 5 and n. cerv. 6. N. cerv.

N. radialis

Arterial branch following radial nerve

N. musculooutaneus *

Cranially coursing branch "Island" in artery

N. medianus Main artery of the limb N. phrenicus Fig. 452.- — Reconstruction of the arteries and nerves of the right arm of a human embryo 9.5 mm. long, viewed ventrally. (After Elze, Anat. Hefte, Bd. 35, Taf. 17-18, Fig. 3.) Some years ago Leboucq (1893) reported that in human embryos of about this age (7 to 11 mm.) the primary vessel of the limb coursed in the forearm between the anlagen of the radius and ulna, and represented the a. interossea volaris. Zuekerkandl had been led to expect this fact by comparative-anatomical studies which indicated that the interossea volaris was the oldest trunk in the lower arm, as well as by embryological observations on other mammalian embryos. His studies, constituting the first researches on the development of these vessels, remain of fundamental value.

Comparative. — Zuekerkandl (1894) thus described the condition of the vessels in the fore limb of rabbit embryos 8.9 mm. long, in which the skeleton was just indicated by mesodermal thickenings. The brachial artery, after accompanying the median nerve in a typical way in the upper ami, is continued in the



forearm as a stem lying next the skeletal anlagen, covered by the flexor pre-muscle mass. Just below the elbow, the artery gives off a branch which goes through to the dorsal side of the forearm (a. interossea dorsalis). As the main artery continues to go distally, the median nerve turns away from it to become superficial, leaving its internal interosseus branch to accompany the axial vessel, which may thus now be called the a. interossea volaris. As the main part of the median nerve leaves the stem artery it is supplied by the latter with an accompaniment of fine vessels which continue with it along its entire superficial course to the palm, where they constitute a superficial palmar plexus. The axial artery itself divides at the distal end of the forearm into a ramus volaris, which breaks up to constitute a delicate deep volar plexus next the skeletal anlagen, and a strong ramus dorsalis which supplies the back of the hand. In rabbits somewhat older the fine vessels

Dorsal branch

Branch accompanyN. radialis N. ulnaris ing the n. radialis Arterial ring


N. cerv. 6

A. subclavia

N. medianus N. musculo- Main artery cutaneus of the limb


Branch directed cranially

Fig. 453. — Vessels and nerves of the same arm (Fig. 452) shown from above. (After Elze, Anat. Hefte, Bd. 35, Taf. 17-18, Fig. 4.) along the main median nerve constitute an artery large enough to begin to dispute the field with the interossea volaris, the a. mediana, while the ulnar nerve has a small accompanying vessel, a. ulnaris. Essentially the same conditions are shown in an 11 mm. cat embryo.

De Vriese has found that in the human embryo of 10 mm. the chief nerve trunks are all accompanied by capillary vessels, 61 and has chosen to represent these as already constituting arterial

81 Injected mammalian embryos show that this is partially true, although the poor material with which De Vriese worked has justly led to the conviction that perineural spaces were confused with true capillaries, a fact which the illustrations accompanying her research leads us to suppose.


pathways. It is doubtful whether these should all be given the valuation which she has set on them, and it must be left to future research to modify or confirm the conception of an already quite complicated arterial system which her description gives us.

The author recognizes at this early stage the a. n. interossea dorsalis, a. n. radialis, a. n. ulnaris, a. n. mediani, and a, n. interossea volaris, the last of which constitutes the continuation of the- axial stem and divides just above the carpus into dorsal and palmar branches, which are themselves in communication by means of a small a. perforans carpi. Four vascular planes are distinguished in the hand, two palmar and two dorsal.

In an embryo measuring 11.7 mm. Miiller has reconstructed the chief arterial system of the limb, and, inasmuch as the nerves and the anlagen of the humerus, radius, and ulna were evident, homologized the vessels present with those occurring in the adult (Fig. 454).




Fig. 454. — Reconstruction of the arteria system of the arm in a human embryo 11.7 mm. long. After Miiller, Anat. Hefte, Bd. 22, Taf. 25-26, Fig. 9.) a. a., a. axillaris; a.b.s.s., a. brachialis superficialis superior; a.b.s.m., a. brachialis superficialis media; a.b.s.i., a. brachialis superficialis inferior; S-, widening of the arterial tube after it has passed through the ventral plexus plate; a.b.s., distal part of the a. brachialis superficialis; a.b.p., a. brachialis profunda; a.r., a. radialis; a.i., a. interossea; a.m., a. mediana; a.u., a. ulnaris; a.a.b.s., a. antibrachii superficialis.

The subclavian perforates the brachial plexus in the usual manner from its ventral side, but the strong branch which, as in previous stages, is given off just before the penetration of the plexus to continue on the medial side of the latter, sends an anastomosing branch through the ventral nerves to join the main vessel. This anastomosing branch joins the main stem at or near the origin of the radial artery from the latter, and on its course toward the chief trunk splits into other branches, as the figure shows. Two of these branches also run into the main trunk, one by piercing the median nerve, the other by going under the same, while another branch courses along volar to the median to anastomose eventually with the median artery branch of the main stem again. In the other limb of the same embryo three strong perforating branches join the part of the main artery medial to the ventral nerve mass with the stem lying lateral to the same. So that in both limbs we are dealing with a rather plexiform condition of the axillary artery. 62 62 Great interest attaches to the arterial plexus formed by the three vessels which penetrate the brachial plexus in this embryo, for it again indicates the employment by the arterial stream of several rather than a single channel from the pre-existing capillary plexus. This transitory plexus axillaris arteriosus of Miiller has also been seen in other mammals (Goppert in the mouse). It by no

DEVELOPMENT OF THE VASCULAR SYSTEM. 665 The main vessel pursues a general course along the radial border of the median nerve to become in the forearm the a. interossea volaris. In its upper-arm portion it gives off, in addition to the a. mediana, a small vessel which joins a chain of capillaries lying in front of the radius anlage (identified by Miiller as the a. radialis). Another branch of the main stem joins correspondingly small vessels lying along the ulnar nerve, and constitutes the a. ulnaris.

In his embryos of 14 mm. Miiller has identified the a. profunda brachii, the a. mediana, a. interossea vol., a. radialis, and a. ulnaris. 63 The main vessels in a 16.2 mm. embryo may be seen at a glance from Fig. 455, in which the lower-arm and hand areas have been omitted.

Mention has already been made of the fact that if, for example, the point of union of the sixth aortic arch with the aorta dorsalis be taken as a fixed point, the subclavian artery appears to wander upward. "Whereas in the embryo of 4.9 mm. the a. subclavia is some eight segmental spaces below this point, in the embryo of 7 mm. it is but six spaces below it, and in the embryo of 11% mm. it is opposite the sixth arch.

The a. mammaria interna and the a. thyreo-cervicalis are conspicuous stems in the embryo of 15.5 mm. (Fig. 447). The a. thyreo-cervicalis courses for some distance in the wall of the jugular lymph-sac in this embryo, and, as McClure has observed the same thing in embryos of the cat, the relation is probably of general significance.

In reviewing the facts hitherto acquired concerning the history of the arm vessels, one must be struck with the need of more careful means constitutes an invariable intermediate stage in the development of the arm vessels, but, as Miiller himself has shown, may occur on one side of the embryo, while the opposite axillary is composed of but a single trunk. Such phenomena may be expected to occur somewhat more frequently in the developing vascular system than in the adult, inasmuch as the increasing blood current exercises a more definite choice to the elimination of multiple paths, but that they may persist even here is shown by the occasional presence of islands in the course of adult arterial trunks.

63 In two 14 mm. embryos Miiller has f ound the main arteiw splitting into two branches which surround the median nerve and fuse again, and is consequently of the opinion that we have here a retention of the arterial paths ventral to the median nerve shown in the previous stage. That this has been the case is strengthened by the fact that the two limbs of the ' brachialis meet again at the place of origin of the a. radialis which is quite an exact correspondence with the place of opening of one of the anastomosing channels shown in Fig. 454. In one of these embryos the artery lying ventral to the median nerve is the larger of the two, while in an older embryo (W. P. 20.5 mm.) it is the persisting one. It seems reasonable to suppose that we are dealing here with instances of a so-called superficial brachial artery, which, as is well known, lies on the volar side of the median nerve, while the normal brachial is dorsal to the latter.


studies here. 64 The reconstructions of Miiller and Elze are our sole possessions in this field. Viewed from a more general stand-point, however, the history of the arm vessels in man certainly confirms the conclusions arrived at some years ago by Zuckerkandl in his studies on the general morphology of these vessels, for in man also the primary artery is an axial stem from shoulder to hand and in the forearm constitutes the later a. interossea volaris. There is also a very general agreement among all observers in the important role played by the embryonic a. mediana. 65 However, there is as complete an agreement in the recognition that at first the volar interosseus is the chief lower-arm vessel, and no support whatever for the idea of Janosik who speaks of the mediana in that primary role.

N. to. A*, u.

Fig. 455. — Reconstruction of the nerves and arteries of the arm in a human embryo 16 mm. long. (After Miiller, Anat. Hefte, Bd. 22, Taf. 27-28, Fig. 6.) The vessel accompanying the radial nerve is the a. profunda brachii (a. nervi radialis of De Vriese). A. a., a. axillaris; A.i., a. interrossea; A.m., a. mediana; A. r., a. radialis; A. u., a. ulnaris; A", m., n. medianus; A", m. c, n. museulocutaneus; N. r., n. radialis; A', u., n. ulnaris.

Comparative. — The a. interrossea volaris with a. perforans carpi is the chief vessel of the forearm in the adult in amphibia, reptilia, and in Ornithorhynchus among the mammalia (Zuckerkandl). It is also apparently the plan in the embryos of all the mammals. (Zuckerkandl, rabbit, cat; Hochstetter, Echidna; Grosser, bats; De Vriese, calf.) In very many mammals the chief definitive arterial stem is the a. mediana (marsupials, edentates, most carnivores, bats, etc.). In the primates its territory is taken over by the ulnar. A vessel accompanying the ulnar nerve, hence an a. nervi ulnaris, occurs in adult amphibia and reptiles and apparently constantly in the embryos of mammals. In the adults of the latter class the artery is not, as a rule, important and may be lacking entirely (most ungulates). A vessel which can be designated the radialis is not of general occurrence till the mammals are reached, and in the majority of these is a superficial radial. It comes to possess its deep volar territory in the higher mammals, but its proximal end is still superficial (really the superficial brachial here) in many of the primates, as Bayer has well shown.

M Very recently Goppert (1910) has supplied us with the history of the development of the arteries in the arm of the white mouse, an account which is by far the most complete we possess for any mammal.

55 The observations of De Vriese even indicate that this vessel is not finally displaced from the hand until the embryo reaches almost 30 mm. in length.

DEVELOPMENT OF THE VASCULAR SYSTEM. 667 Variations. — Many of the variations of the arm vessels must remain uncertain in origin until we possess a well verified series of observations on their embryology. There can be no doubt, however, that Miiller has demonstrated the manner in which a superficial brachial may arise, for arterial channels are retained on the ventral side of the median nerve in most of his specimens. It may be pointed out, also, that eases of persistence of great median or even volar interosseus arteries (Baader) are unquestionably survivals of embryonic conditions, and we may have all possible degrees of variation in the part taken by these vessels in the supply of the hand (Schwalbe and others). Krause pointed out that high origins of the radial or ulnar arteries usually involved a superficial course for the proximal part of these vessels, a fact which may be explained by the retention of a brachialis superficialis inferior. Attention may also be called to the very ingenious series of schemata which Miiller lias constructed to explain the lower-arm arterial anomalies, but until more is learned of the normal history here, we can not venture to present satisfactorily founded diagrams for anomalies.

Arteries of the Lower Limb. — In human embryos measuring from 5.5 to 7 mm. and shortly after the umbilical arteries have acquired their secondary, more lateral, stems of origin from the aorta in the neighborhood of the fourth or fifth lumbar segments, there can be seen going out from these vessels on either side, a small artery which penetrates the tissues of the posterior limb bud (Fig. 420). When the nerve-plate for the lower limb grows out farther, it surrounds this vessel, so that the extremity artery appears now to pass through it, just as is the history with the subclavian artery and the brachial plexus. Later the ischiadic nerve joins this vessel and it may consequently be identified as the a. ischiadica.® 6 Probably injections of earliest stages here would show that the a. ischiadica is really only the exaggerated member of a series of vessels, which originally supply the limb tissue, as is the case with the upper limbs.

This vessel (a. ischiadica) forms a central or axial nourishing channel for the early leg bud, just as is the case with the subclavian and early arm bud. Leboucq (1893) first called attention to the fact that the primitive blood supply of the hind limb consisted in a single axially-coursing artery, the a. ischiadica, which, as soon as skeletal elements could be recognized, continued to course in the lower-leg region between the anlagen of tibia and fibula, and ended chiefly as a strong branch which perforated the interspace between the elements of the first tarsal row, to reach the dorsum of the foot. Lately De Vriese 67 has confirmed this.

88 It is to be noted that in mammalian embryos, where the history of the leg vessels has been followed more carefully, the a. ischiadica is primarily a branch of the aorta, and its proximal portion serves later as the stem of origin for the umbilical arteiy when the latter abandons its ventral roots (Hochstetter, 1890). The origin of the a. ischiadica from the aorta has not yet been observed in man.

87 De Vriese has considered the history of the leg vessels in man. I will not, however, detail her account, for reasons given above in the account of the arm vessels.


It may be well to refer here to the important previous observations of Zuckerkandl (1894-95), who described the leg arteries in a rabbit embryo of 7.7 mm. somewhat as follows : The a. ischiadica is continued in the lower leg as a strong axial vessel next to the skeleton. It sends two perforating branches towards the side of the limb, the upper of which probably corresponds to the a. tibialis ant., while the lower supplies the dorsum of the foot. The distal end of the axial vessel supplies the depths of the sole. Fine vessels accompany the posterior tibial nerve in its lower course, and in embryos of 13.5 mm. these constitute a distinct artery, a branch of the axial vessel. The further history in this animal disclosed the a. saphena (from the a. femoralis) taking over the posterior tibial trunk.

The supremacy of the a. ischiadica in the supply of the extremity is soon disputed by the appearance of a new vessel, the a. femoralis, which in the embryo of 15.5 mm. (Fig. 447) is already the chief vessel in the limb. The femoral soon gains all the branches of the a. ischiadica in the territory of the lower leg {'e.g., tibialis posterior et anterior), by anastomosing with the ischiadica near the knee; we know the a. ischiadica of the adult only as the stem portion of the a. glutea inferior. 6S This ontogenetic history of primary and secondary vessels for the human leg is closely paralleled by the vessels found in an ascending vertebrate series, as Zuckerkandl showed.

Comparative. — The a. ischiadica is the chief vessel of the thigh in the adult for amphibia, reptilia, and the birds, yet the femoral in the latter class may attain quite an area of distribution, and in some {e.g., Spheniscus.) even behaves as in mammals by taking over the chief rami of the a. ischiadica and constituting the chief limb vessel (Hochstetter). On the other hand, among the mammalia the atrophy of the ischiadica is the ride. Yet it may persist in part, as, for example, forming the a. tibialis anterior of bats (Grosser, 1901).

The appearance of an a. ischiadica in the embryos of all mammals was indicated by the observations of Hochstetter and Zuckerkandl. The chief lower-leg portion of the a. ischiadica in adults in the amphibia, reptiles, and birds behaves exactly as it does in the early stages of the embryo of man, namely, courses between the tibia and fibula and supplies the dorsum of the foot by means of a large perforans tarsi. Other stages in the ontogeny of man's leg vessels are found definitive in various mammals, e.g., the stage in which a distinct superficial plantar arch or plexus exists, as well as a deep one. This is the case in most apes, as Popowsky has shown, and is lost in the anthropoids, where, as in man, the lateral plantar artery is larger than the medial and the deep plexus practically the only one present.

Much interest attaches to the saphenous artery. The earlier work of Zuckerkandl emphasized the very general occurrence of this vessel in all the mammalia/' 9 and led him to declare it the oldest (phylogenetically speaking) branch of the 68 According to Hochstetter's (1S91) investigations on mammals the a. comes n. ischiadici does not appear to be a relic of the old ischiadica, although this assumption is made by most authors. De Yriese describes the lower leg portion of the original axial artery (a. n. interossei cruris) as becoming the a. peronea of the adult, giving over all of its important branches in the territory of the foot to the a. tibialis anterior.

19 With the exception of Bradypus bidaetylus, Lemur catta, and man, in which it is much atrophied.

DEVELOPMENT OF THE VASCULAK SYSTEM. 669 femoral. In its lower portion this artery usually takes over the dorsalis pedis artery, or the primary tibialis posterior, or both, in which latter case it constitutes the chief or 011I3- vessel for the supply of the foot. The vessel retains its importance in the primates, — e.g., in Cebus, where it supplies the entire foot. Popowsky has recalled the anomalous occurrence of this vessel in man and reported two interesting cases in which the a. saphena was large, in both cases anastomosing with the posterior tibial artery and in one case in addition with the dorsalis pedis. He has again called attention to the frequent great development of this vessel in the monkeys, where, even in the anthropoids, it supplies the dorsalis pedis. Popowsky, evidently much influenced by this, states his belief that this vessel must play a prominent role in the development of the leg arteries of man. There is no evidence, however, that such is the case. The work of De Vriese indicates there is apparently no necessity for the recapitulation of a stage in which the saphenous functions as the chief artery of the lower leg. In the reworking of this field, nevertheless, great interest will attach to the re-examination of the embryonic importance of this vessel, for the reasons above given.

Variations. — Dubrueil, Krause, Ruge and others have described cases in which the a. ischiadica was the chief vessel of the limb in man, which is quite evidently a survival of embryonic conditions. The occurrence of an a. saphena magna, following the saphenous nerve, has already been mentioned, the first case having been observed by Zagorsky in 1809. Normally this vessel probably reaches the lower third of the leg, for in well-injected subjects I have traced it this far, as Hyrtl first did. Krause and more recently Salvi report cases where an artery accompanies the n. cutan. suras lat. This corresponds to an embryonic vessel seen by De Vriese at the peroneal side of the leg in the 13 mm. embryo, but it usually disappears entirely. Cases in which the a. peronaea instead of the tibialis ant. supplies the dorsum of the foot are not rare, and represent again the embryonic picture where the axial artery behaves normally thus. Most interesting are cases in which a perforans tarsi persists, joining the dorsalis pedis with the deep plantar vessels. In the adult also a superficial plantar arch occasionally occurs, as Krause, Gegenbaur and others mention.

D. The Development of the Veins.

1. The venous types.

2. The ground-plan of the young venous system.

3. Transformations of the vv. umbilicales et vitelline.

4. Transformations of the vv. cardinales anteriores.

5. Transformations of the w. cardinales posteriores.

6. The development of the veins of the body walls.

7. The development of the veins of the extremities.


In the adult, as has been recognized for a long time, the veins tend everywhere to follow the arteries, — i.e., the majority of the veins are w. comites. In the embryo, however, it is possible to satisfy one's self that this is not the primary arrangement, for. if one studies carefully the developing vessels in any area, it will be seen that the earliest arterial and venous trunks are separated from one another so as to stand in reciprocal relation as regards the general capillary bed. Should this primary separation of


arteries and veins be perpetuated as the vascular trunks continue to grow, we have the plan which obtains, for instance, in the circulation of the- brain or lung, where larger arterial and venous vessels instead of coursing together are arranged so as to stand opposite one another. As a rule, however, as development proceeds the main vascular stems are found coursing together, — i.e., the veins are true w. comites. We have to recognize, then, tivo types of veins, primary and secondary veins, primary veins standing opposite or alternating with the arteries and trunks, secondary ones coursing in company with the corresponding arteries. 70 Venae comites, which are, then, always later formations, may arise either as a result of shifting of primary trunks in growth or entirely de novo. 11 Splendid examples of the persistence of primary veins are furnished by the great subcutaneous veins of the limbs and trunk (v. basilica, v. saphena, v. thoraco-epigastrica). These are in fact remains of the early border veins of the extremities and of very early body wall trunks and it may hence appear more reasonable why they possess no corresponding accompanying arteries.


If now one turns to the details of the developing venous system in man, it will be recalled that the remarkable precocious development of the chorionic circulation gives us the vv. umbilicales at stages much earlier than obtain in the mammalia generally. In embryos of 6 somites (N.T. 3) we can also trace clearly the vv. vitellines, and it can be seen that in their terminal portions the umbilical veins join the heart by receiving the vitelline veins and coursing now as a common vitello-umbilical trunk. 12 At the margins of the anterior intestinal portal, this vessel turns inward, courses in the mesial wall of the pleuropericardial passage, and in the mesodermic tissue ventral to the foregut anastomoses with its fellow of the opposite side to constitute the sinus venosus. The latter is at first situated in front of the first somite (Mall embryo 391, with seven somites), but in the fifteen somite embryo (Graf Spee No. 52) is opposite the body of the first somite, and in the twenty- three somite embryo (N.T. 7) is opposite that of the sixth (Thompson, 1908). In this last stage there open into the sinus

70 Even though their peripheral portions, of course, always exhibit a primary separation from the arteries.

71 The primary circulation schema and the secondary birth of venae comites may be seen beautifully in such an expanded flat area as the area vasculosa of the chick (cf. Popoff), but no less clearly, for example, in the extremities, where the primary border vein drains all the blood from the central artery, whereas secondarily venae comites arise (Hochstetter, 1891).

2 This common vitello-umbilical vein of man corresponds really to the end portion of the vitelline veins of other early mammalian embryos in which always umbilical veins secondarily appear later (e.g., rabbits).

DEVELOPMENT OF THE VASCULAR SYSTEM. 671 the anterior and posterior cardinal veins by means of a ductus Cuvieri, but earlier, when the sinus lies more crarfialward, the anterior cardinal vein joins the common vitello-umbilical vein (embryo of fifteen somites). The intermediate stages are not known in man.

At twenty-three somites, then, we have present the four pairs of veins (the vv. cardinales anteriores et posteriores, the vv. umbilicales, and the w. vitelline), which form an entirely symmetrical venous ground-plan, characteristic not only for man but for all the vertebrates. This ground-plan of the venous system remains in embryos which are approximately a centimetre long, and its existence in man has been known to us since the classical descriptions of His (1880 to 1885).

It will be convenient to study the development of the adult venous tree as a modification of each of these primitive systems. The proximal ends of the umbilical and vitelline veins enter into special relations with one another in the region of the liver, and with the further growth of the liver bud are converted into the two venous trees of that organ, the vv. hepaticae and vv. porta?. On account of the early inauguration of these changes, they may be described first.


Mention has already been made of the fact that the vitelline veins are first interrupted in their course to the sinus venosus by the growth of the liver bud, which in embryos from three millimetres on in length, begins to cause the interposition of many smaller vessels (sinusoids of Minot) in the venous current through the liver. The early stages in this process can be seen in the figures supplied us by His (Figs. 425 and 426), where both vitellines have as yet a fairly direct path through the liver region and open on either side into the sinus venosus. Very soon, however, the left v. omphalomesenterica is more effectively cut up into nourishing liver capillaries (sinusoids), although these still drain into the left horn of the sinus venosus by way of the old opening there of the original vein, which hence constitutes a primitive v. hepatica sinistra (Fig. 456).

This persists as late as in embryos of 7 mm. (Elze).

The umbilical veins next gain connections with the liver sinusoids and eventually lose completely their early superficial course in the region between liver and heart, a fact first noted by H. Rathke (1838). 73 The umbilical blood is now poured into the liver channels, the largest of which is the old direct path of the right omphalomesenteric to the corresponding horn of the sinus venosus. In the mean time the two vitelline veins have anastomosed with

Cited after His, Anat. Mensch. Embryonen III, S. 210 (1885).

672 each other by cross connections, which go, first ventral, then dorsal, and again ventral, to the gut tube and so form two venous rings around the duodenum, as may be seen from Fig. 456. 74 The middle or dorsal of these anastomoses receives the vein from the intestine, the true mesenteric vein.

His pointed out that the usual fate of these venous rings involved always the atrophy of certain limbs and the persistence of others in such a way that an S-shaped course is now described

Ductus venosus Arantii

Junction between the ductus Arantii and the end piece of the left yolk vein

Sinustvenosus, right.Lorn

V. omphalo mesenterica dextra

V. umbilicalis dextra

Sinus venosus, left horn

V. omphalomesenterica sinistra

Cranial anastomosis of the vitelline veins

Middle anastomosis of the vitelline vein

Right vitelline vein

Caudal anastomosis of the vitelline vein

Left vitelline vein Fig. 456. — Schema of the liver circulation in a human embryo 4.9 mm. long (NT. 14). (After Ingalls, 1907.

by a common vitelline trunk in reaching the liver. It is important to note that during this time the left umbilical vein has effected a direct connection with the cranial venous ring and that the right umbilical atrophies. The right vitelline vein also disappears, so that by the time the embryo is 7 millimetres in length the main source of blood for the liver comes from the left vitelline and left umbilical veins. 75 The liver end of the former vessel is the old S-shaped common vitelline trunk, and where it becomes dorsal to the gut, consequently the place corresponding to the early dorsal venous anastomosis, — the middle one of the three, — it receives the 74 The researches of Hochstetter make it probable that these venous rings (first seen by His in the human embryo) are of very general occurrence among the mammalia.

i5 The umbilicalis dextra still connects with the liver sinusoids in the 7 mm. embryo (Elze).



mesenteric vein. Somewhat further headward and after it has turned around the right side of the gut to become ventral to it, and at a place corresponding to the former cranial venous ring, this, now the omphalomesenteric trunk, receives the left umbilical vein. For a time the chief channel for all this blood through the liver is the intrahepatic course of the former right vitelline vein (Mall) (Fig. 458). Soon the development of an anastomosis (already indicated in Fig. 456) enables the vena hepatica sinistra to lead its blood also into the right end of the sinus venosus, near the opening of the right vitelline trunk (secondary v. hepatica sinistra), while the former multiple afferents of the left omphalo

Venae revehentes

Ductus venosus Arantii Pane. d. Venae revehentes

Cranial end of the v. umbiliealisdextra .

Stomach - -_

Vense advehentes

Ductus choledochus V. umbilicalis dext

Cranial end of the v. umbilicalis sinistra

_,.-^g ^---- Vena? advehentes

Li ver

V. umbilicalis sinistra

Obliterated part of Duo- V. ntellina V. vitellina An. v. c. Obliterated part of the venous ring denum dextra sinistra the %-enous ring Fig. 457. — Schema of the liver circulation in the human embryo at a later stage than that shown in Fig. 456. (After His, from Marshall.) Pane, d., pancreas dorsale; An. v. c, annulus venosus caudalis.

mesenteric into the left lobe of the liver are now reduced to a single larger supplying trunk, the ramus angularis (Mall).

When, with the growth of the right lobe of the liver, the intrahepatic course of the right vitelline becomes shifted so as to constitute a somewhat circuitous route, a new direct one to the sinus venosus is formed; this is the ductus venosus Arantii. Mall's researches show that the former intrahepatic course of the right vitelline does not completely atrophy without a trace, but leaves representatives in the form of its end, which drains into the sinus venosus, and its first portion, which leaves the umbilical vein, for these are now incorporated as parts of the supplying (portal) and draining (hepatic) systems of the liver, and become respectively the ramus dexter vena? hepatica? and the ramus arcuatus (et descendens) vena? portae. At this stage, then, we have for both of Vol. II. -43'



flit- two main divisions or Lobes of the liver, porta] or supplying and hepatic or draining trunks; on the left, the ramus angularis venas portae, the blood from which is drained into the ramus sinistra venae hepaticae, on the right, the ramus arcuatus of the portal vein, opposite to which stands the ramus dexter of the hepatic (Mall) Fig. 459).

In an embryo 11 mm. long two trunks have been added to both the supplying and draining systems, and four units or lobes may be described as being present. To the portal system have been added the right and left arborizations of the recessus nmbilicalis (Eex, 1888), to the hepatic the ramus medius and vena cava inferior (Fig. 460). Xowthe middle and left hepatic veins both

I io ~< inidiagrarnmatie reconstruction of the veins of the liver of a human embryo 5 mm. long, Mall No. 80. ) (After Mall, 1906.) L., liver; u. v., umbilical vein; v. o. m., right omphalomesenteric vein; r.h.s., ramus hepaticus sinister; r. u., reces3us umbilicalis; r. a., ramus angularis; m., mesenteric vein; /., intestine.

Fig. 159. — Semidiagxammatic reconstruction of the vein?- of the liver of a human embryo 7 mm. long, Mall Xo. i. (After Mall, 1906.) /...liver: u. v., umbilical vein; m., mesenteric vein; r.u., recessus umbilicalis; d. v. ductus venosus; r. a.,. ramus arcuatus; r.h. </., ramus hepaticus dexter; r. h. «., ramus hepaticus sinister.

divide, and consequently by the stage of 26 millimetres we find six collecting trunk.-, the upper and the lower right hepatic (the latter a branch of the inferior cava), the right and left media, and the ii] 'per and lower left hepatics. Correspondingly six supplying trunks exist, for the right portal branch splits into a ramus ascendens as well as ramus dexter, and, in addition to the ramus angularis, we have also the left arborization of the recessus umbilicalis and two other prominent brandies of this trunk, one of which may be identified as it- right arborization (Fig. 461). Mall pointed out that these six primary lobules of the liver correspond with the six lobes to be recognized in the morphology of the adult mammalian liver (Bex).

ft ha- been pointed out that at the stage of 4.9 mm. the dorsal anastomosis between the vitelline veins receives the mesenteric vein draining the intestine. After the 8-shaped common vitelline vein



is formed out of these anastomoses and after the right vitelline vein lias atrophied, the left vitelline becomes the sole efferent from the yolk sac and receives the mesenteric 'vein at the earlier point of union of the latter with the dorsal anastomosis. Prom here on to the liver then this vein is properly the omphalomesenteric vein, hut in most of its conrse it has been free from the mesentery, crossing the coelome Independently of the latter. On the other hand, the omphalomesenteric artery, which supplies both gut and yolk-sac, courses in the mesentery. The artery is directly trans

. A I-"u;. 160 Reconstruction of the vascular system of the li\i-r of a human embryo 11 mm. long. i Mall No. 109.) I \itiT Mall, 1906.) u. v., umbilical vein; p. v., portal vein; r.a., ramus annularis; r. «., ras umbilicalis; r.d., ramus descendens; r. a., ramus arcuatus (possibly ramus ascendens); r.c, rinht arborisation of the recessus umbilicalis; r. /., left arborisation of the recessus umbilicalis; J. v., ductus venosus; i c, vena cava;, omphalomesenteric vein; r. m., ramus medius; r. 8., ramus sinister.

formed into the superior mesenteric artery, hut its accompanying vein (v. mesenterica superior) is a secondary channel which has arisen to drain the gut wall and it alone, the yolk sac drainage going by way of the former left vitelline vein. Only a small pari of the vitelline vein is incorporated in the vena porta o\' the adult, namely, thai part proximal to where the mesenteric vein is re ceived. 78 " It was Luschka (1863) who first pointed out that the vitelline vein does not persist in the v. mesenterica superior, although tins Is largely true for the corresponding artery. Dexter (1902) mid Lewis (1903) for the eal and pig, and Bonnol and Seevers (1906) in the case "t 1 man, have called specific attention to this Pact.



The anterior cardinal vein suffers profound modifications, for it and its derivatives come to form the sinuses of the dura while its proximal portion constitutes the great internal jugular trunk of the adult. We have already seen that in human embryos of 15 somites the anterior cardinal or, better, the primitive head vein can be identified from the region of the fore-brain to its opening into the common vitello-umbilical vein opposite the third somite, and that it can be divided into a longer portion lying- in front of the region of the somites and a shorter portion in the segmental

Fig. 461. — Reconstruction of the vascular system of the liver of a human embryo 2-1 mm. long(Mall No. 6.) (After Mall, 1906.) u. v., umbilical vein; v. p., vena portse; r. u., recessus umbilicalis; r. a., ramus arcuatus; r. d., ramus descendens; r. a., ramus angularis; r. c, right arborization of the recessus umbilicalis; r. 1., left arborization of the recessus umbilicalis; v. h., vena hepatica; d. v*, ductus venosus; d. s., vena dextra superior; d. i., vena dextra inferior; m. d., vena media dextra; m. s., vena media sinistra; 8. s., vena sinistra superior; s. i., vena sinistra inferior; v. c, vena cava.

area ; the former portion lies close at the sides of the hind-brain and should be known as the v. capitis medialis (Grosser, 1907) ; the latter is situated more laterally and is the true cardinalis anterior. 77 Both portions of the primitive head vein are in fre 77 Grosser first separated these two portions of the primitive head vein, which occur in all vertebrates, and called attention to the fact that only the caudal part is homologous with the posterior cardinal and hence merits the name cardinalis anterior. He remarks that the cardinals are probably especially related to the segmental excretory system and that the anterior cardinal is likely evidence of the former cephalic extent of this.



quent connection with the aorta by means of numerous small direct branches, the v. capitis medialis by means of dorsal presegmental arteries, the true cardinalis anterior by means of the dorsal segmental arteries as well as by direct lateral branches of the aorta. Later all of these aortic offshoots atrophy, and the chief source of the blood drained by the primitive head vein is supplied by the a. carotis interna. When the anlagen of the cranial nerves first appear, they are found lateral to the vena capitis medialis, but in later stages, as Salzer (1895) first showed, the vein is gradually shifted lateral to the nerves by the formation of channels which course on the outer side of the latter, and the v. capitis lateralis thus produced gives us a secondary, wholly lateral, head vein.

Fig. 462. — Reconstruction of the veins of the head in a human embryo 9 mm. long.

(After Mall, 1905.)

(Mall No. 163/

The development of vascular sprouts which enable the medial head vein to begin to circumvent the ganglia of the cranial nerves occurs early. In embryos 3 mm. in length (Broman, NT. 11) it has shifted lateral to the acustico-facialis, the otic vesicle, and the glosso-pharyngeus. This position we saw it had retained in the embryo of 4.9 mm. (NT. 14). When the sixth nerve can be identified, it also is medial to the vein. Next the tenth nerve is surrounded by a venous ring ami the lateral path around this nerve chosen, to the elimination of the medial oneSuch a ring around the vagus may be seen in 7 mm. embryos (Fig. 420) or. again. may not be formed when a length of 9 mm . is reached (Fig. 462). Gradually a similar loop forms around the Gasserian ganglion (Fig. 463). From the fifth nerve caudalward to the twelfth, then, the medial head vein has become the v. capitis lateralis.

n The v. capitis medialis iu the region of the fifth nerve is retained to become the sinus cavernosns of the adult (Mall), but otherwise the early medial head vein leaves no trace of its existence. The v. capitis lateralis is entirely without the skull, or, more accurately, leaves the skull with the seventh nerve to empty ts blood into the internal jugular vein, and so it takes no part in the formation of the permanent head sinuses, although its chief tributaries do so, as Mall has shown in the following way. At the stage of which we are speaking, the v. capitis lateralis possesses three main tributaries, the anterior, middle, and posterior cerebral veins respectively (Mall). The first of these drains the eye (v. ophthalmica) and cerebral hemispheres as well as mid-brain; its most cephalic extension conrses on either side of the mid-dorsal

Fig. 463. — Reconstruction of the veins of the headin a human embryo 11 mm. long.

(After Mall, 1905

Mall No. 109. 1

line in the region of the fore-brain and constitutes the anlage of the erior sagittal sinus, thus primitively paired. The middle cerebral vein drains the anterior part of the hind-brain (cerebellum) and joins the main trunk between the fifth and seventh nerves. Since the v. capitis lateralis leaves the skull in company with the seventh nerve, it is apparent that through this foramen the venous blood of the fore-brain, mid-brain, and cerebellum is drained. The last tributary of the lateral head vein joins it behind the otic vesicle, leaving the skull through the embryonic jugular foramen (v. eerebralis posterior). This posterior cerebral vein drains the remainder of the hind-brain (medulla) and first portion of the cervical cord. As the anterior cerebral vein extends forward to the top of the cerebrum, so also the posterior cerebral reaches the mid-dorsal region of the hind-brain (Pig. 464).

Xow anastomoses develop between these three primitive cerebral veins and the v. capitis lateralis atrophies, so that not only

DEVELOPMENT OF THE VASCULAR SYSTEM. 679 the hind-brain blood but that of the entire brain is drained out through the foramen jugulare, and the old anterior exit with the n. facialis disappears. 78 The anastomoses which develop between these three primitive brain veins begin the changes that convert these to the head sinuses. The blood from the sinus sagittalis superior is no longer returned by way of the anterior cerebral vein, but courses dorsally by means of a new anastomosis which links it to the upper end of the cerebralis media. Very soon, though, an anastomosis is carried still further caudally, so that the blood now enters the posterior cerebral vein, which leaves the skull through the jugular foramen. This last and most important anastomosis forms the

Fig. 464. — The right vena cerebralis posterior (Mall) draining the roof of the hind-brain in a human embryo 11 mm. long. (Mall No. 353.) Injection preparation. (After a sketch kindly placed at my disposal by Mr. Max Broedel.) major portion of the lateral sinus, and in the fetus of 33 mm. is a large channel which has completely supplanted the old v. capitis lateralis. This great channel is gradually shifted backward in later stages by the growth of the cerebral hemispheres. The 78 Salzer and Mall call attention to the fact that in all probability Kolliker mistook this exit of the v. capitis lateralis from the skull as a drainage of the early head by the external jugular vein, and hence thought that he had confirmed Luschka, who believed that this was the case and that only secondarily did the internal jugular drain the brain. Luschka fancied that the foramen jugulare spurium, to which he first called attention, represented this primary exit of the skull drainage. The internal jugular, however, is from the first the only vein of the brain, and this is true also after the skull begins its development. The external jugular vein is an entirely secondary channel much later to develop. It is of interest to note that Hochstetter has shown that in the adult of Echidna the blood of the anterior part of the brain is drained by the persisting part of the v. capitis lateralis, which leaves the skull with the facialis and thereafter joins the internal jugular trunk. In Ornithorhynchus also, as Hochstetter has shown, the same vein exists, but it is only supplementary here to the vein traversing the foramen jugulare.


original cerebraiis media is probably incorporated to form the superior petrosal sinus, but the inferior petrosal sinus is a later formation. 79 " 1 The v. jugularis externa is a secondary venous channel which in man, as in the mammals generally, appears relatively late (embryo of 16 mm., F. T. Lewis, 1909; see also Schawlowski, 1891). We possess as yet no connected history of the vein for man.

The reader will find the mention of some stages in the development of this vein in the guinea-pig given by Salzer (1895) and in the bat by Grosser (1901).

We have seen that in early embryos the floor of the branchial region is drained on each side by a vein which originally joinsi the duct of Cuvier but is soon a tributary of the anterior cardinal. 79 " Lewis (1909) has traced this vein in a series of embryos, and believes it can be recognized as the linguo-facial vein of the adult, where it usually belongs to the external jugular trunk. Its transfer from the internal to the external jugular appears after the stage of 16 mm.

The proximal ends of what were originally the anterior cardinal veins do not continue to open into the heart separately, — i.e., by means of two ducts of Cuvier, formed by the union of anterior and posterior cardinal veins on each side. Only the right opening persists, and this is possible by the development of a great anastomosis between the anterior cardinals (Fig. 478) which enables the left vein to conduct all its blood into the right one. The anastomosis becomes the v. anonyma sinistra, 80 and that portion of the right anterior cardinal between the opening of the v. anonyma sinistra and the right subclavian vein is known as the v. anonyma dextra, whereas the lower portion of the right anterior cardinal and the right ductus Cuvieri becomes the vena cava superior. The original portion of the left anterior cardinal below the transverse anastomosis becomes the end portion of the v. hemiazygos accessoria, the remainder of which is constituted by the left posterior cardinal ; of the left ductus Cuvieri only the proximal portion is preserved as the sinus coronarius (Marshall, 1850).

79a Grosser (1907) has shown this to be homologous with the inferior jugular vein of fishes.

79b The reader is referred to the recent study of the development of these veins made by J. Markowski (1911). (Markowski, Ueber die Entwieklung der Sinus durae matris und der Hirnvenen bei mensehliehen Embryonen von 15.5—49 mm. Scheitel-Steiss lange, Bull, de l'Acad. des Sciences de Craeovie, Juillet, 1911.) so Schawlowski (1S91) and Anikiew (1909), from fragmentary observations on human embryos, conclude that veins draining the thymus gland are concerned in the formation of this anastomosis (v. anonyma sinistra).


We have seen that the posterior cardinal veins form two long symmetrical drainage channels which receive dorsally segmental afferents si (vv. intercostales et lumbales) and ventrally many small tributaries from the Wolffian bodies, and that, when the hind limbs develop, their chief afferent — the fibular border vein — also opens into the posterior cardinal.

Gradually, now, two veins arise to assist in the drainage of the mesonephros. These are the w. subcardinales (F. T. Lewis, 1902), and have already been noted in the preceding accounts of several embryos (vide pp. 604, 612). They lie on the ventral surface of the mesonephros on each side, and each of them is not only connected at either end and at many other points with the corresponding posterior cardinal vein, but also joins its fellow of the other side by means of cross anastomoses across the front of the aorta. The latter communications are soon confined to one large connection just below the origin of the a. mesenterica superior and at the level of the future w. renales.

Although for a time the subcardinal veins can only thus be considered accessory and tributary to the posterior cardinals, the right subcardinal acquires another highly important connection headward with the vascular system of the liver (the hepatic half), 82 and it is afterward possible for a great part of the blood from the hind end of the body to stream directly into the heart by means of the common hepatic vein (v. hepatica revehens communis). 83 This connection inaugurates a profound change in the drainage of the legs and lower trunk, the end result of which is the substitution of a single large channel — the vena cava inferior — in place of the earlier multiple and symmetrical veins.

For the details of this change we are indebted mainly to the investigations of F. Hoehstetter and of F. T. Lewis on the rabbit. A complete account for man,

81 It may again be emphasized that in the beginning the posterior cardinals receive more of the cervical segmental veins than later. These, with the exception of the first, drain into the v. cardinalis posterior, but with the descent of the heart and great vessels, the cervical veins become tributaries of the anterior cardinal.

82 It is of interest to note that Davis (1910) has demonstrated open connections between the subcardinal veins and the portal system in early embryos of the pig, but these reach their maximum and are obliterated before the vena cava is formed.

83 Hoehstetter thus names the trunk passing from the liver to the heart and formed, as we have already seen, from parts of the hepatic, umbilical, and omphalomesenteric veins. It has been pointed out that some of the blood from the lower limbs and tail can stream through the sinusoidal vessels of the Wolffian body and join the vena cava, thus giving us a partial renal-portal system for the mesonephros of mammals. Yet in mammalian embryos we must grant Hochstetter's remarks that the characteristic renal-portal system of Sauropsida is only approached.

682 founded on a satisfactory series of human embryos, is still lacking. I shall accordingly content myself here with a brief presentation of the essential facts won from other mammalian embryos and of the probable history in man. This is the more justifiable also, since we possess many scattered observations, on such human material as has been at hand, by Hochstetter, Zumstein, and Kollmann, among others.

The exact manner in which the cardinal system is tapped by the hepatic was pointed out by F. T. Lewis (1902) and more recently by D. M. Davis (1910). The latter observer has shown that the capillaries on the ventral surface of the Wolffian body proliferate in a cephalic direction, fusing with capillaries which surround the oesophagus (peri-cesophageal plexus) and which course also on the

V. hepatica comm.


Cava mesenterii

Capillaries on the end of the v. subcardinalis

V. portse

Anastomoses between Anlage of the Anlage of the dorsal Aorta ventral pancreas pancreas

Anastomoses oetwei the umbilical veins

Fig. 465. — Sagittal section through a pig embryo 8 mm. long, showing the hepatic and subcardinal capillaries approaching one another to form the vena cava inferior. (After Davis, 1910.)

wall of the stomach. Thus the drainage territory of the subcardinal vein is extended headward. On the right side, beyond the anterior limit of the Wolffian body, this skirmish line of capillaries grows in the connective tissue of the caval mesentery which has also been invaded by hepatic capillaries in advance of liver cells. Soon hepatic and subcardinal capillaries meet and fuse, and for the first time a vascular path is offered from the right subcardinal to the common hepatic vein (Fig. 465). Inasmuch as both subcardinal and cardinal veins are in frequent connection, this new path diverts much of the blood stream of the lower posterior cardinal, which formerly went to Cuvier's duct, through this new channel. Thus in the posterior cardinal veins we may now be said to have two blood streams, for the current in the lower part of both veins turns ventrally into the upper right subcardinal vein by virtue of the great anastomoses between cardinals and sub

DEVELOPMENT OF THE VASCULAR SYSTEM. 683 cardinals, whereas in that part of the posterior cardinals above the level of the Wolffian bodies the blood goes upward to the ductus Cuvieri. This leads to a more or less complete separation of the two portions of the posterior cardinal vein. The upper portions of these veins are transformed into the system of the azygos and hemiazygos veins of the adult; the lower portions undergo still other changes. 84 For a while, although disturbed by the migration of the permanent kidneys, 85 they remain quite symmetrical, and so the vena cava appears double in the region below the great anastomosis above mentioned. 86 Eventually, however, only the lower segment of the right posterior cardinal persists to constitute the peripheral segment of the single adult vena cava inferior, for the left vein atrophies 87 in virtue of anastomoses between the two cardinals which enable the right channel to drain satisfactorily all the blood. The chief of these anastomoses (the transverse iliac vein) enables the blood from the left pelvic region, and the left limb to drain practically entirely into the right cardinal. In this way the transverse iliac vein constitutes the tenninal portion of the left common iliac, which has hence a morphological value different from the terminal part of the right v. iliacus communis. 88 Anastomoses also enable the left lumbar veins to be carried across the vertebral column to open into the right lower cardinal (cava), whereas the upper great anastomoses between the cardinals remains as the proximal part of the left adult renal vein. It is only necessary to add that the subcardinal veins below the level of the great transverse anastomoses atrophy, while that portion of the left vein above this level functions as the proximal part of the left adrenal vein. It hence goes into the renal vein (which represents in part the original great trans-anastomoses), rather than into the vena cava directly, as the right adrenal vein does. • The vena cava inferior, then, is a composite vessel, and is formed, from the liver downward, of parts of the following veins : right hepatic vein, connecting vein in the caval mesentery, right upper subcardinal vein, and right lower posterior cardinal.

84 These lower portions of the posterior cardinal veins persist symmetrically in some mammals and so form a vena cava which is double below the level of the vv. renales (Echidna, Edentates, Cetacea).

85 As the anlage of the permanent kidney ascends from the pelvis to its permanent position, it appears to push in between the aorta and the posterior cardinal vein and to displace the latter ventral- and lateralward. A more direct collateral venous path is developed going dorso-medial to either the ureter or the kidney anlage, which may for a time be thus surrounded by a venous ring. (Vide Hochstetter, 1S93; Zumstein, 1887 and 1S90; Grosser, 1901; Lewis, 1902.) 84 An arrangement which may persist in those well-known anomalies in which we have a double cava below the kidneys.

OT Hochstetter states that the lower left cardinal obliterates up to the point of reception of the spermatic (ovarian) vein, and that consequently the end portion of this lower left cardinal is represented in the most proximal part of the left spermatic vein of the adult. The opinion that part of the left cardinal is represented by the ascending lumbar vein (Lewis, Bryce) is disputed by him, on the ground that the latter has a more lateral position. He assigns the origin of this vein to secondary anastomoses which establish a chain between the thoracic and iliac region. It is of interest to note that the atrophy of the left lower cardinal is not the only method by which a single adult cava is produced in the region below the kidneys. In some mammals this is attained apparently by a true fusion of the two cardinals dorsal (Ornithorhynchus) or ventral (most Marsupials) from the aorta (Hochstetter).

  • "Which is probably only the proximal part of the early v. ischiadica.

684 The upper portions of the posterior cardinal veins are undoubtedly concerned in the formation of the vv. azygos and hemiazygos. 89 Here again, though, we possess as yet no accounts for the embryo of man. The arrangement of the veins in question in the adult shows that normally in further growth an asymmetrical development of these two veins occurs. This, nevertheless, is not



V. I. f.

. u.

Fig. 466. — Injection of a pig embryo 8 mm. long, showing the extensive system of transverse bodywall tributaries to the umbilical vein. (After Smith, 1909.) V.l.f., vena linguo-facialis; S. r., sinus reuniens; Pars sup. v. r., pars superior v. umbilicalis; m. r., membrana reuniens; v. u. d., vena umbilicalis dextra; V. c. a., v. cardinalis anterior; V. c. p., v. cardinalis posterior.

usually so extreme as is the case, for instance, with the rabbit, where the right vein alone persists. In man, as is well known, the left trunk is only interrupted, for, while the lower portion joins 88 In all accounts hitherto given us, the upper portions of the original posterior cardinal veins have been described as entirely separated from their lower portions by the great deflection of the venous blood current due to the appearance of the inferior cava, and this "separation" occurs at such a level (e.g., the eighth thoracic segment, rabbit) that these upper portions of the posterior cardinals must be subsequently extended to the end of the thoracic region to constitute the a2ygos of the adult. They are, in fact, described as actually "growing down" secondarily. Hochstetter (1903, p. 604) comments on the conditions he found in a 15.5 human embryo, in which the adrenal glands destroyed the symmetry of the posterior



the right vein (v. azygos) by means of one or more large cross anastomoses, its upper portion, the so-called v. hemiazygos accessoria, continues to Cuvier's duct. 90 Information on the exact details of the transformations effected in these venous channels in various mammals should be sought in the papers of Hochstetter, Zumstein, Lewis, Grosser, McClure, Gosset, Parker and Tozier, Van Pee, Beddard, Soulie and Bonne.

Plexus v. e. s.

Fig. 467. — Injection of a pig embryo IS mm. long, showing the superficial body wall veins. (After .Smith, 1909.) Plexus v. e. s., plexus of superficial epigastric vein; V.m.i., v. mammaria interna; Y.t.e-, v. thoraco-epigastrica.


We have seen that in young embryos the body walls are drained into the umbilical vein by an extensive svstem of tribu

eardinals: "Audi hat dieses Organ (die Nebenniere) das Kopfende der Urniere, welches sich somit schon sehr stark retrahiert hat, so weit lateralwiirts abgedrangt, class ein Zusammenhang der v. azygos und hemiazygos niit don Venen dieses Organs nicht mehr bestehen kann. Der geschilderte Befund lasst bedeutende Zweifel dariiber aufkommen, ob die v. azygos und hemiazygos beim Menschen in ihrer Totalitat als Reste der hinteren Kardinalvenen aufzufassen sein werden." 90 But exceptional cases in which an entirely symmetrical doubled schema ipreserved are by no means uncommon in man, and in some mammals, on the other hand, this is a normal course of development, — e.g., Echidna (Hochstetter).

686 taries. There is no doubt, then, but that we must regard the v. umbilicalis as the primary drainage channel for the body wall. 91 Its domain here is next disputed by the appearance of the v. thoraco-epigastrica? 2 which forms on the lateral body wall just caudal to the arm bud. Proximally, the thoraco-epigastrica unites with the primitive ulnar vein to constitute the v. subclavia, which, as Hochstetter (1891) first showed, at first courses dorsal to the brachial plexus and subclavian artery to enter the v. cardinalis anterior (embryo of 10 mm., F. T. Lewis, 1909), but in slightly

Fig. 468. — Injection of a pig embryo 15 mm. long, showing symmetrical mid-ventral veins draining the plexus situated in the membrana reuniens over the heart.

older embryos possesses also an opening ventral to these structures, so that in the latter stage (embryo of 11.5 mm., F. T. Lewis, 1909) the a, subclavia and plexus brachialis are enclosed in a venous ring, only the ventral limb of which will persist.

91 Since the complete system of these veins has not yet been figured for human embryos, I present here three figures to show their extent in another mammal (the pig). Miss Smith's figures (Figs. 466, 467) have been secured from injections of living embryos, and I supplement them by a figure to show the plan of mid-ventral drainage (Fig. 468). Here one remarks that the membrana reuniens over the upper portion of the heart territory is drained by two parallel mid-ventral veins which eventually join the v. umbilicalis. (In some instances they also end by branches which sink in directly to the vessels of the liver.) 12 Homologous with Hochstetter's "Seitenrmnpfvene" of the lower vertebrates.



Owing to the fact that at first the lateral body walls greatly exceed in extent the dorsal and ventral surfaces, their chief drainage channels, the vv. thoraco epigdstricce, are the most important body-wall veins until relatively late (embryo of 50 mm). What proportion of the body-wall drainage they still control in an embryo of 35 mm. can be seen from Fig. 473. At this later stage, however, the more ventrally lying veins begin to play a significant role, among which are to be mentioned the superficial epigastrics and the perforating branches of the vv. mammaria internee and inter co stales. In embryos of 50 mm, injections show that the territories of these latter veins have grown very appreciably, yet there do not occur as yet any appreciable anastomoses, such as produce here the great venous plexus well known in the adult (Fig. 472).


We lack as yet any thorough-going account of the development of the extremitv veins in man. Nevertheless, the researches of

P. C. V

T. e. v,

F g. 469. — Injected human embryo 11 mm. long, showing some of the chief superficial veins. (From a drawing by Mr. Max Broedel.) (Alall No. 353.1 T.e. v., thoracoepigastric vein; P. c. v., posterior cerebral vein; U- v., umbilical vein.

F. Hochstetter (1891) on the extremity veins of Amniotes and the scattered observations which have been made on the human embryo, together with some others which will be presented here, enable us to outline the essential facts in this field.

The first veins of the limb bud in man, as in other mammals and in the chick, are small direct vessels which drain the early capillary plexus of the limbs into the posterior cardinal and umbilical veins. These venules thus constitute two sets — a dorsal series, which are the tributaries of the posterior cardinal vein,


V. fetnoralis superficial

Fig. 470. — Injected human embryo 20 mm. long, showing some of the chief superficial veins. (Mall No. 349. (After drawings kindly placed at my disposal by Mr. Max Broedel.)

Fig. 471. — Injection showing the thoraco-epigastric and superficial epigastric veins in a human embryo 35 mm. long. (Mall No. 449.) Fig. 472. — The same in an embryo 50 mm. long. (Mall No. 458.) The relative growth of the lower .vein is evident. No anastomoses between the two systems are yet present.

and a 'ventral series, the tributaries of the umbilical vein. Such are the conditions in human embryos under five millimetres in length.



Ramus perforans v. mammaria? interna

V. thoraco epigastrica

nus perforans v. mammalia? interns

Rami perforantes v. mammaria? interna?

V. thoraco. epigastrica

exus venosus mammilla;

Ramus perforans v. mammaria internae

V. epigastrica superficialis

V'. femoralis superficialis externa

V. saphena magna \V. saphena parva — ^^_ (v. saphena accessoria) Vv. dorsalis penis cutanea? Fig. 473. — Body-wall veins of a human fetus 35 mm. long. (Mall No. 449,) The specimen was secured alive through the kindness of Dr. Thomas Cullen and injected through one of the aa. umbilicales.

But there is soon established in both limbs (in the anterior limb first and later in the posterior) a border vein which surrounds the paddle-like extremity, 93 a vein which Hochstetter has shown to be characteristic for the limb bud of all the amniota. The

93 The observations of Lewis and Grosser have indicated that both radial and tibial border veins are extremely transitory; Grosser, in fact, was not able to find a tibial border vein in the bat ; however, Bardeen figures this clearly in his studv of the leg bud of a 11 mm. human embryo ( Amer. Jour. Anat.. I, 1901. PI. IV, Fig. D, p. 36).

Injections of the limb buds of pig einbryos show that the border vein is constructed out of the peripheral margin of the capillary plexus of the limb.

Vol. II.— 44

690 upper (radial and tibial) portions of these border veins are quite insignificant, but the lower (ulnar and fibular) ones are relatively large 94 and constitute the chief channels of drainage of the extremities. Moreover while the radial and tibial border veins completely atrophy, the ulnar and fibular veins persist, their peripheral portions constituting the basilic and small saphenous veins of the arm and leg respectively. Proximally the ulnar border vein constitutes the definite branchial, axillary, and subclavian vein. For a considerable time this is the only important venous channel in the arm, and, although its proximal portion still functions as the





Y.scl.d. \ %-V.cardposi

Figs. 474 and 475. — Reconstructions of the veins of the right arm in two human embryos 10 and 11.5 mm. long respectively. (After F. T. Lewis, 1909.) V.card.ant., vena cardinalis anterior; V., v. cardinalis communis; V .card. -post., v. cardinalis posterior; V.ling.fac, v. linguo-facialis; V.scl.d., v. subclavia dorsalis; Vscl.v., v. subclavia ventralis;, v. thoraco-epigastrica; V.ul.p., v. ulnaris prima.

chief vein in the adult limb, its distal superficial territory is soon greatly exceeded by the development of the v. cephalica.

In embryos of ten millimetres and under, the proximal portion of the ulnar border vein, after receiving the thoraco-epigastric vein from the lateral body wall, drains into the posterior cardinal or common cardinal vein by taking a course dorsal to the brachial plexus and subclavian artery (Fig. 474, F. T. Lewis). Shortly after this stage, however, a venous path is also found ventral to these structures, and after a short time, during which the brachial nerves are enclosed in a venous ring (Fig. 475), the dorsal path finally For some reason the limb capillaries will not approach very close to the ectodermal covering of the limb bud, but leave a narrow sub-ectodermal zone of mesenchyme non-vascular; hence the marginal vein which is formed from the " frontier line " of these capillaries follows faithfully the boundary of the rim.

84 Hochstetter observed in living embryos that the direction of blood flow for practically the entire extent of the border vein of the upper limb is from before backward, i. e., into the ulnar extremity.



W^SSk~ '"card.ant.

Fig. 476. — Reconstruction of the veins of the right arm in a human embryo 16 mm. long. (After F./T. Lewis, 1909.) V. ceph., v. cephalica. For other abbreviations see Figs. 85 and 86.


•Sty '-an. sin. =9fij V.mam. int.


Fig. 477. — Reconstruction of the right shoulder region in a human embryo 22.8 mm. long. (After F. T. Lewis.) Ribs, clavicle, scapula, and humerus have been stippled and the subclavius muscle has been drawn. V. an.dext., v. anonyma dextra; V. an. sin., v. anonyma sinistra; V. br., v. brachialis; V. ceph., v. cephalica; V. jug. ant., V.jug.ext., V. jug. int., v. jugularis anterior, externa, et interna; v. mammaria interna.

692 atrophies. Moreover, while the subclavian vein at first opens into the posterior 'cardinal, it eventually is found joining the duct of Cuvier, and in still older embryos (16 mm.) the anterior cardinal or jugular rein, a phenomenon to he associated with the descent of the heart and main vessels into the thorax.

The cephalic vein is entirely secondary, and appears first in man, as in the rabbit (Fig. 469), ° 5 as a small vessel which collects the blood from the outer side of the hand plate and fore-arm anlage and flows into the radial end of the ulnar border vein near the elbow. 96 Very soon this vein can be traced upward along the

Y. jugulai. int. f -.

V. jugularis ext.

V. jugulo-cephaliea

Anastomoses between tne \ v. anonvms

V. jugulocephalica


V. cepha lica

V*. cephalica

V. brachialis

V. brachialis

V. basilica

r \

V. basilica

Y. thoraco-epigastrica

Y. thoraco-epigastrica

Fig. 47S. — Reconstruction of the relations of the great veins of the arms and neck in a human embryo 20 mm. long. (Mall collection, No. 349.) radial side of the upper arm (Figs. 476, 477), and in an embryo of 22.8 mm. Lewis has shown that the cephalic vein now joins the external jugular, an arrangement which is true for the embryo figured in Fig. 478, but in which there is also now present a connection between the cephalic and the subclavian veins which is to function as the definitive proximal ending of the cephalic vein in man. This earlier drainage channel of the cephalic into the external jugular vein may persist (jiigulocephalic rein), as has been noted for many years in descriptive human anatomy.

The cephalic vein at the stage last mentioned has become the chief superficial vein of the arm, for, with the breaking up of the

  • > Compare with Hochstetter's figure 2 a, Taf . III. Morpli. Jahrb., 1891, for the rabbit.

" This connection of basilic and cephalic veins has nothing to do with the v. mediana cubiti, which is a late connection and formed long after the primitive junction of the two vein? has disappeared and they have existed as two independent channels. (See beyond.)



border vein by the outgrowth of the digits and the formation of interdigital veins, we have a transferral of the latter veins to the system of the v. cephaliea, which now, collecting its blood from the back of the hand, courses along the radial border of the forearm and arm entirely distinct from the ulnar border vein (the v. basilica, Fig. 479). As is well known, in the adult these two great veins are connected in a wide-meshed plexus. A complete injection

Jg£- V. cephaiica

,Vv. intercapitulares V. mediana antibrachii

V. basilica

I' [G8. 479 and Iso. — The superficial veins of the right arm in a human fetus 35 mm. long. From an injection. (The specimen is the same as that shown in Fig. 473.) of the arm veins in an embryo 35 mm. long shows thai even at this stage there are not yet formed the many connections between basilic and cephalic veins which constitute the well-known venous plexus of the dorsum mani and the forearm. It is thus possible to state that the great subcutaneous venous plexuses of the extremity are not partial remains of a primary embryonic more

694 extensive plexus, for the only primary plexus existing here is again a general capillary mesh, and the larger venous connections which characterize the adult are clearly secondary formations. In the arm figured, one may see the earliest veins of the volar surface of the forearm, and, especially clearly, the method of formation of the v. mediana antibrachii through the enlargement of parts of the general capillary mesh (Fig. 480).

In the posterior limb bud it has already been mentioned that the superficial portion of the fibular border vein persists, for it can be identified in a series of embryos (15.5, 20, 23, and 26 mm. long) and seen to constitute the v. saphena parva. The

Fig. 481.

Fig. 482.

Fig. 481. — The fibular border vein in a human embryo 15.5 mm. long. (Mall collection, No. 390.) (After a sketch kindly placed at my disposal by Mr. Max Broedel.) Fig. 482. — The fibular border vein in an injected human embryo 21 mm. long. (Mall collection, No. 460.) The vein is seen to drain the dorsum of the foot by a distinct venous arch; the proximal portion of the original border vein can be recognized.

deep portion of this vein accompanies the sciatic artery and nerve in the region of the thigh and through the foramen ischiadicum into the pelvis ; it is hence the v. ischiadica. It joins the posterior cardinal vein, of which it constitutes the chief radicle, for the caudal vein (v. sacralis media) is inconspicuous. At a later stage (Fig. 485) the vein formed from the union of the femoral and great saphenous veins joins the proximal portion of the ischiadic vein just before the latter ends in the v. cardinalis posterior. In human embryos measuring 10 mm. or less, the ischiadic vein constitutes the chief drainage channel of the lower limb, but in its superficial extent the vein is soon exceeded by the v. saphena magna, a secondary channel, and in its deep territory by



the v. femoralis, which has developed along the permanent artery (a. femoralis) of the limb (the v. ischiadica in the adult being important only as a collateral path for the blood). The early

V. saphena parva

Fig. 4S3. — The fibular border vein (v. saphena parva) in a human embryo 23 mm. long (Mall No. 462) at a time when toes and heel are clearly evident.

Fia. 484. — Drainage of the perineum and buttocks into the v. saphena magna, in a human fetus 50 mm long. (Mall No. 458.) (From an injection by Mr. Broedel.) development of the v. saphena magna in man is not known, but at the stage of 23 mm. it already constitutes the chief superficial vein of the leg.

In embryos of 24 and 25 mm. length, anastomoses on the inner side of the thigh have begun to direct the blood stream in the

696 saphena parva to the v. saphena magna, and in an embryo measuring 35 mm. and in three embryos of approximately 50 mm. in

V. cava inf.

V. epigastrica Vena femoralis superficialis superficialis externa

V. saphena magna

V. saphena parva

Vv. saerales media? Fig. 4S5. — Reconstruction of the chief veins of the pelvis and lower extremities in a human embryo 20 mm lone. (Mall collection, No. 349.)

V. femoralis superficialis externa

V. saphena magna

V. saphena parva (v. saphena accessoria) Fig. 486. — The superficial veins of the leg in a human fetus 35 mm. long. (After an injection of the living embryo; secured through the kindness of Dr, Thos. Cullen.) (Mall, No. 449.)

length, I have found this connection a constant feature, practically all the blood of the lower leg vein {v. saphena parva) going



into the greater saphenous channel. 97 In the youngest of these embryos the saphena parva continues up the inner side of the thigh before joining the saphena magna (a condition which lias heen observed as a variation in the anatomy of the adult for a long time), but in all the other cases the small saphenous vein pours its blood into the v. saphena magna near the knee. Eventually the v. saphena parva joins the deep vein {v. femoral is) in this neigh

V. .-aphena accessoria

V. saphena magna

Fig, 487. — The superficial veins of the leg in a human fetus 50 mm. long. (After an injection by Mr. Broedel.) borhood, as is well known to be its definite normal ending, although in a great percentage of cases the connection here with the saphena magna is also retained to form a subsidiary channel {e.g., Quain's Elements of Anatomy, 10th Ed., Vol. II, Part II, p. 538, London 1894.)

"Whether Ave are dealing here with a general fact or not is impossible as yet to decide. If such is not the case, it must be remarked as unusual that I have found the six lower limbs of the three embryos measuring fifty millimetres to be absolutely indentical in this respect. I note also that Bardeleben refers to a .similar arrangement of the saphenous veins. " Ferner mundel bei jenen (d. h. Feten) die v. saphena parva, welche der basilica homolog ist, in die saphena magna" (Bardeleben, 1880, p. 604).


In the development of the leg, the proximal portion of the extremity is for a while buried, as it were, in the tissues of the embryo, and only in embryos of some 20 mm. in length, and in those older than this, can we speak of a cutaneous surface belonging to the inner side of the thigh. Consequently the saphena parva is in the position to drain the early venules which come from the neighborhood of the perineum and buttock (if we may yet speak of the latter), as Fig. 481 will show. With the "pushing out" of the thigh, this is no longer possible, 9s for the proximal end of the saphena parva is carried out with the knee, and the saphena magna is now the direct and natural channel for this blood. In embryos measuring 50 mm. in length the vessels draining the back of the buttocks into the saphena magna constitute a large and prominent system (Fig. 484).

As Bardeleben first indicated and as has been shown by the work of Hochstetter and of Lewis, the limb veins which are true accompanying vessels to the arteries are the last to develop." BIBLIOGRAPHY.


Aeby, C. : Der Bau des menschlichen Korpers. Leipzig 186S.

Arnold, J. : Experimentelle Untersuchungen iiber die Entwicklung der Blutkapil laren. Virchow's Arch. Bd. 53, S. 70. 1871. Backman, G. : Ueber Inselbildung im Gefasssvstera. Anat. Hefte. Bd. 38. Heft 114. S. 1. 1909. Balfour, F. M. : Comparative Embryology. London 1885.

Billroth, T. : Untersuchungen iiber die Entwicklung der Blutgef asse. Berlin 1856. Bobritzky: Entwicklung der Kapillargefasse. Centralbl. f. med. Wiss. Bd. 23.

1885. Broman, J. : Beschreibung eines menschlichen Embryo. Morph. Arb. Bd. 5. 1S96. Bryce, T. H. : "Embryology" in Quain's Anatomy, 11th Edition. London 1908. Clarke, E. R. : Observations on Living Growing Lymphatics in the Tail of the Frog Larva. Anat. Record. Vol. 3, No. 4. 1909. Drasch, O. : Die Bildung der Somatopleura und der Gefasse beim Hiihnchen. Anat.

Anz. Bd. 9. 1894. Duval: Etudes stir Fembryologie des Cheiropteres. Journ. de l'Anat. et de la Phvs.

1897. Elze, C. : Beschreibung eines menschlichen Embryo von ea. 7mm. grosster Lange usw. Anat. Hefte. Bd. 35. Hefte 106. S. 411. 1907.

"* Except, of course, in those eases in which, as in Fig. 4S6, the saphena parva continues up the inner side of the thigh.

99 Mention may here be made of the occurrence of a prominent " superficial external femoral " vein which drains the front and outer side of the thigh and joins the saphena magna near the fossa ovalis (embryos of 23 and 35 mm.). It is relatively large in these embryos, by far the greatest tributary of the saphena magna. The vein is recognized as common in the adult {e.g., Quain, Spalteholz, Piersol, etc.). In the 35 mm. embryo it was remarkable that on both sides the vein began suddenly by deep roots in the region of the knee — the latter streamed from the vascular plexus surrounding the cartilage of the lower end of the femur and turned out sharply to the skin.

DEVELOPMENT OF THE VASCULAR SYSTEM. 699 Eternod, A. C. F. : Communication sur un oeuf humain avec embryon excessive ment jeune. Xleme Congres medical de Rome. Section d'Anatomie. 1894. Sur un oeuf humain de 16mm, 3 avec embryon de 2,11 mm, uterus et annexes.

Actes de la Soc. helvet. des sc. natur. p. 164-169. 1896. Premiers stades de la circulation sanguine dans l'ceuf et l'embryon humain. Anat.

Anz. Bd. 15, S. 181. 1898. II y a un canal notochordal dans l'embryon humain. Anat. Anz. Bd. 16, S.

131-143. 1899. L'anse veineuse vitelline des Primates (Homme et Quadrumanes). C. R. de P Association des Anat. 4eme Sess. p. 103-110. Montpellier 1902. Zwei menschliche Embryonen von 1.3 mm und 2.11 mm Lange usw. Gynaeco log-ica helvetica. 4. Jahrg. S. 21-23. 1904. L'ceuf humain. Memoire public a l'occasion du Jubile de l'Universite de Geneve.

Geneve 1909. Evans, H. M. : On the Development of the Aorta, Cardinal and Umbilical Veins, and the other Blood-vessels of Vertebrate Embryos from Capillaries. Anat.

Record. Vol. 3, p. 498-518. 1909. Fleischmann: Embryologische Untersuchungen. Heft 1 u. 2. 1889-1891. Frassi, L. : Ueber ein junges menschliches Ei in situ. Arch, f . mikr. Anat. Bd. 70, S. 492. 1907. Weitere Ergebnisse des Studiums eines jungen menschlichen Eies in situ. Arch.

f. mikr. Anat. Bd. 71, S. 667-694. 1908. Gerard, G. : Les canaux anastomotiques arterio-veneux chez Fhomme et chez les singes. Arch, des Sci. Medicales. T. 1 und Les anastomoses directes entre les arteres et les veines. Poid. T. 3. 1896. Goette, A. : Entwicklungsgeschichte der Unke. Leipzig 1875. Golubew: Beitr. zur Kenntnis des Baues und der Entwicklung der kapillaren Gefasse. Arch. f. mikr. Anat. Bd. 5, S. 85. 1869. Grosser, O. : Zur Anatomie und Entwicklung des Gefasssystems der Chiropteren.

Anat. Hefte. Bd. 17. 1901. Ueber arterio-venose Anastomosen an den Extremitatenenden beim Menschen und den krallentragenden Saugetieren. Arch. f. mikr. Anat. Bd. 60, S.

191-216. 1902. Heape: The Development of the Mole. Quart. Journ. of Microsc. Scien. N. S.

Vol. 23. 1883. Hess, W. : Eine mechanisch bedingte Gesetzmassigkeit im Bau des Blutge fasssystems. Arch. f. Entwicklungsmechanik. Bd. 16. 1903. His, W. : Anatomie menschlicher Embryonen. Heft 1~3. Leipzig 1880-1885.

Lecithoblast und Angioblast der Wirbeltiere. Abh. d. math.-phys. Klasse d.

Kgl. Sachs. Ges. d. Wiss. Bd. 26. Leipzig 1900. Hochstetter, F. : Die Entwicklung des Blutgef asssystems. Hertwig's Handbuch der vergl. u. expt. Entw. d. Wirbeltiere. 1906. (Erschien 1903.) Houssay: Developpement et morphologie du parablaste et de l'appareil circula toire. Arch, de Zool. exper. et gener. T. 8. 1890. Hubrecht, A. A. W. : Early Ontogenetic Phenomena in Mammals and their Bearings on our Interpretation of the Phylogeny of Vertebrates. Quart. Journ.

of Microsc. Scien. Vol. 53. 1908. Die Saugetierontngenese in ihrer Bedeutung f. d. Phylogenie d. Wirbeltiere.

Jena 1909. Ingalls, N. W. : Beschreibung eines menschlichen Embryo von 4.9 mm. Arch, f .

mikr. Anat. Bd. 70. 1907. Janosik: Le developpement des globules sanguins chez des Amniotes. Biblio graphie anatomique. T. 10. 1901. Keibel: Die Entwieklungsvorgange am hinteren Ende des Meerschweinchenem bryos. Arch. f. Anat. u. Phys. 1888. Ein sehr junges menschliches Ei. Arch. f. Anat. u. Entw. 1890.


Keibel und Elze: Nomientafel zur EntwickluDgsgeschiehte des Mensehen. Jena.

1908. Kxower: Effects of Early Removal of the Heart, etc., Anat. Record. Vol. 1. 1907. Kolliker, A. : DevelopjDement des tissus chez les Batrachiens. Annales des Sci.

natur. 3d Ser. No. 6. 1S46. Entwickslungsgesehiehte des Mensehen und der hoheren Tiere. Leipzig 1879. Die Entwicklung' der Keimblatter des Kaninchens. Festschrift. Leipzig 1882. Kollmax, J.: Die Kdrperform menschlicher normaler unci' pathologischer Ernbry onen. Arch. f. Anat. u. Phys. Anat. Abt. Suppl.-Band. S. 105-137. 1SS9. Lehrbuch der Entwieklungsgeschichte des Mensehen. 1898. Handatlas der Entwieklungsgeschiehte des Mensehen. 1907. Lewis, F. T. : The Intra-embryonic Blood-vessels of Rabbit Embryos from 8V2 to 13 Days. Amer. Journ. of Anat. Vol. 3, No. 1, p. 12-13. 1904. The Question of Sinusoids. Anat. Anz. Bd. 25. S. 261. 1904. Lexer, E. : Entstehung entziindlieher Knochenherde und ihre Beziehung zu den Arterienverzweigungen der Knochen. Arb. d. ehir. Klin. Bd. 17. Berlin 1904. Lillie, F. R.: The Development of the Chick. New York 1908. Lockwood, C. B. : On the Development of the Arteries of the Abdomen and their Relation to the Peritoneum. Proc. Roy. Soc. Vol. 38. London 1884. Low. A. : Description of a Human Embryo of 13-14 Mesodermic Somites. Journ.

of Anat. and Phys. Vol. 42. p. 237. 1908. Mackay, J. J.: The Arterial System of Vertebrates Homologically Considered.

Memoirs and Memoranda in Anatomy. Vol. 1. 18S9. Mall, F. P.: A Human Embryo Twenty-sis Days old. Journ. Morph. Vol. 5, p. 459-480. 1891. The Development of the Human Ca?lom. Journ. Morph. Vol. 12, p. 395.

1897. Report upon the Collection of Human Embryos nt the Johns Hopkins University. Anat. Record. Vol. 5, No. 7. 1911 . M~khxert, E.: Leber Ursprung und Entwicklung- des Hamovasalgewebes. Morph.

Arb. Bd. 6. 1896. aIixot. C. S. : Human Embryology. New York 1892.

On a Hitherto Unrecognized Form of Blood Circulation without capillaries, in the Organs of Vertebrates. Proceedings of the Boston Society of Natural History/ Vol. 29, p. 185-215. 1900. Laboratory Text-Book of Embryology. Philadelphia 1903. 2d Edition 1910. Mollier, S. : Die Entwicklung der Gefasse im Embryo. Handb. vergl. u. ex)>t.

Entwickl. d. Wirbeltiere, herausg. von O. Hertwig. Bd. 1, Kap. 5, S. 1261.

1906. Oppel, A.: Leber die gestaltliche Anpassung der Blutgefasse unter Beriicksiehti gung der funktionellen Transplantation. Mit einer Originalbegabe von W.

Roux. Vortr. u. Aufs. lib. Entwieklungsmechanik. No. 10. Leipzig 1910. Piper: Ein menschlicher Embryo von 6.S mm. Nackenlange. Arch. f. Anat. u.

Phys. 1900. Rabaud, E. : Sur l'origine endodermique des vaisseaux sanguins. C. R. Soc. Biol.

Ser. 10, T. 3. 1896. Rabl, C. : Uber die Bildung des Herzens bei Amphibien. Morph. Jahrb. Bd. 12.

1SS7. Raffaele: Sistema vascolare nei Selaeei. Mit. Zool. Stat. Neapel. 1892. Raxvier, L. : Du developpement et de l'accroissement des vaisseaux sanguins. Arch.

Phys. norm, et path. Vol. 1, p. 429. 1S74. Remak: Ueber blutleere Gefasse im Schwanze der Fi'oschlarve. Arch. f. Anat. u Phys. S. 79. 1S50. Rexaut, M. J.: Sur la Variation modelante des vaisseaux sanguins. Compt. rend.

de L'assoc. des Anat. Vol. 3, p. 63, 1901. and vol. 4, p. 230. 1902.

DEVELOPMENT OF THE VASCULAR SYSTEM. 701 Robixsox : Observations upon the Development of the Segmentation Cavity, etc.

Quart Journ. of Microsc. Scien. N. S. Vol. 33. 1892. Rouget: Memoire sur les capillaires sanguins et lymphatiques. Arch. d. Phys. p.

603. 1S73. Roux, W. : Uber die Verzweigungeli der Blutgef asse. Jenaische Zeitschr. f. Naturw.

Bd. 13. 187S. RiiCKERT u. Mollier: Die Entsteh'ung der Gefasse mid des Blutes bei Mirbeltieren.

Handb. d. vergl. u. expt. Entwicklungslehre d. Mirbeltiere, herausg. von 0.

Hertwig. Bd. 1, S. 1019. 1906. Ruge, G. : Beitrage zur Geiasslehre des Menschen. Morph. Jahrb. Bd. 9. 18S3. Schulte, M. vox, u. Tilxey, F. : A Note on the Organization of the Venous Return. Anat. Record. Vol. 3, p. 555. 1909. Spee, Graf : Ueber einen menschlichen Embryo usw. Mitt. f. d. Verein Schleswig Holst. Arzte. Heft 11. 1SS7. Beobachtungen an einer menschlichen Keimscheibe mit offener Medullarrinne und canalis neurentericus. Arch. f. Anat. u. Phys. Anat. Abt. S. 159. 1SS9. Neue Beobachtungen uber sehr friihe Entwieklungsstufen des menschlichen Eies. Arch. f. Anat. u. Phys. Anat. Abt. 1S96. Stricht, C. vax der: L'origine des premieres cellules sanguines et des premiers vaisseaux sanguins dans l'aire vaseulaire de chauves-sauris. Bull, de l'Acad.

•Roy. de med. de Belgique. T. 13, p. 4. 1S99. Stricker, S. : Untersuchungen uber die kapillaren Blutgef iisse in der Xiekhaut des Frosches. Sitz.-Ber. d. kais. Akad. d. Miss. Mien 1S65. Studien liber den Bau und das Leben der kapillaren Blutgef iisse. Sitz.-Ber. d.

kais. Akad. d. Miss. Mien. Bd. 52. 1866. Taxdler, J.: Zur Entwicklungsgeschichte der arteriellen Mundernetze. Anat.

Hefte. Bd. 31. 1906. Thoma, R. : Untersuchungen fiber die Histogenese und Histomechanik des Ge fasssysteins. Stuttgart 1S93. Thomsox, A. : On the Development of the Vascular System in the Foetus in Verte brated Animals. 2 parts. Edinb. New Phil. Journ. 1830-1831. Thompsox, P.: Description of a Human Embryo of Twenty-three paired Somites.

Jour, of Anat. and Phys. Vol. 16. 1907. Turstig, J. : Untersuchungen fiber die Entwicklung der primitiven Aorten. Sckrift., berausgegeben v. d. Xaturf.-Ges. bei d. Univ. Dorpat. Bd. 1. Dorpat 1SS4. Van Beimelen, E. and Julix. C. : Recherches sur la formation des annexes fcetales chez le mammiferes. Arch. d. Biol. T. 5. 1SS4. Vastarixi, C. G. : Le anastomosi artero-venose nell' uomo e nei mammif eri. Studio anatomo-Lstologico. p. 176. Napoli 1902. Vialleton, L. : Developpement des aortes chez l'embiyon de poulet. Journ. de l'Anat. et de la Phys. T. 28, p. 1. 1892. Vosmaer, G. C. J. : On the Retrograde Development of the Blood-vessels in the Omentum of the Rabbit. Verslag v. d. Gew. Vergad. d. Mis. Naturk. Afd.

Deel 6, p. 245. 1S9S. MiEDERSHEnr, R.: Vergleich. Anatomie der Mirbeltiere. Fiinfte Auflage, Jena.

1902. Der Bau des Menschen. Vierte Auflage. Tubingen 1908. Youxg, A. H. : The Development and Morphology of the Vascular System in Mammals. Studies ia Anat. from the Anat. Dept. Owens Coll. and Journ.

Anat. and Phys. Vol. 32. 1900. Zimmermaxx : Rekonstruktion eines menschlichen Embryo von 7 mm Lange. Verb.

d. anat. Ges. 3. Vers. Erg.-Heft z. Anat. Anz. 1889.


Aebt, C. : Der Bronchialbaum der Saugetiere und des Menschen. Leipzig 1SS0. Allex. M. : Omphalomesenteric Remains in Mammals. Journ. of Anat. and Phvs. Vol. 17. 1883.


Baader, A. : Ueber die Varietaten der Armarterien des Menschen und ihre morpho logisehe Bedeutung. Inaug.-Diss. Bern 1866. Baer, K. E. v.: Ueber die Kieinen und Kiemengefasse in den Embryonen der Wirbeltiere. Meckel's Arch. 1827. Des branches et des vaisseaux branchiaux dans les embryons des animaux vertebres. Annal. des Sciences Naturalist. Tome 15. 1828. Barpi, U. : Intorno ai rami minori del' aorta abdOminale ed all' irrigazione arteriosa del ganglio semilunare, del plesso solare e delle capsule surrenali negli equini, nei carnivori e nei roditore domestici. Arch. Ital. Anat. e di Embriol. Vol.

1, S. 491-522. 1902. Bayer, F. : Beitrag zur vergleichenden Anatomie der Oberarmarterien. Morph.

Jahrb. Bd. 19. 1893. Bemmelen, J. F. van: Die Visceraltasehen und Aortenbogen bei Reptilien und Vogeln. Zool. Anz. Bd. 9. 1886. Boas, J. E. V. : Ueber Herz und Arterienbogen bei Ceratodus und Propterus.

Morphol. Jahrb. Bd. 6. 1880. Ueber den conus arteriosus und die Arterienbogen der Amphibien. Morphol.

Jahrb. Bd. 9. 1883. Ueber die Arterienbogen der Wirbeltiere. Morphol. Jahrb. Bd. 13. 1887. Bremer, J. L. : On the Origin of the Pulmonary Arteries in Mammals. Amer.

Journ. of Anat., vol. 1, p. 137-144, 1902 ; and Anat. Record, vol. 3, No. 6, p. 335. June 1908. Description of a 4 mm. Human Embryo. Amer. Journ. of Anat. Vol. 5, p.

459. 1906. Broman, I.: Die Entwicklungsgeschichte der bursa omentabs. Wiesbaden 1904. Ueber die Entwicklung, Wanderung und Variation der Bauchaortenzweige bei den Wirbeltieren. Ergeb. der Anat. u. Entw. Bd. 16. 1906. Tiber die Existenz eines embryonalen Pfortaderkreislaufes in der Nachniere der Saugetiere. Anat. Anz. Bd. 31. 1907. Uber die Entwicklung und "Wanderung" der Zweige der aorta abdominalis beim Menschen. Anat. Hefte. Bd. 36, S. 407-549. 1908. Coulter, C. B. : The Early Development of the Aortic Arches of the Cat, with Especial Reference to a Fifth Arch. Anat. Record. Vol. 3, p. 578-592.

1909. Castellani, L. : Osservazione sullo sviluppo della circolazione sanguigna del rene umano. Ricerche fatte nei Labor. Anat. normale di R. Univ. di Roma et in altri Labor. Vol. 12. Fasc. 4, p. 225-252. 1907. Clark, John G. : The Origin and Degeneration of the Blood-vessels of the Human Ovary. Johns Hopkins Hospital Reports. Vol. 9. 1900. Dedekind, F. : Beitrage zur Entwicklungsgeschichte der Augengefasse des Menschen. Anat. Hefte. Bd. 38. Heft 114. S. 1. 1908. Dohrn, A. : Studien zur Urgeschichte das Wirbeltierkorpers. XV. Mitt. a. d. Zool.

Stat, zu Neapel. Bd. 9. 1889-1891. Dubrueil: Des Anomalies arterielles. Paris 1847. Engel, E. : Lo sviluppo dei vasi sanguigni nella palpebre dell' uomo. Ric. fatte nei Labor. Anat. Normale di R. Univ. di Roma et in altri Labor. Vol. 12.

Fasc. 4. 1907. Ernst, P. : Unpaariger Ursprung der Intercostal- und Lumbalarterien aus der aorta. Zeitschr. f. Morph. u. Anthrop. Bd. 1, S. 495. 1899. Evans, H. M. : On an Instance of Two Subclavian Arteries in the Early Arm Bud of Man. Anat. Record. Vol. 2. 1908. On the Earliest Blood-vessels in the Anterior Limb Buds of Birds and their Relation to the Primary Subclavian Artery. Amer. Journ. of Anat. Vol 9.

1909. Fedorow, V. : Ueber die Entwicklung der Lungenvene. Anat. Hefte. Bd. 40, S.

529. 1910.

DEVELOPMENT OF THE VASCULAR SYSTEM. 703 Felix, W. : Zur Entwicklungsgeschichte der Rumpfarterien des menschlichen Embryo. Morph. Jahrb. Bd. 41. Heft 4. S. 577. 1910. Fere, C. : Vaisseaux omphalo-rnesenteriques. Bull. Soe. Anat. Paris 1878. Fetzer : Ueber ein dureh Operation gewonnenes menschlicbes Ei, das in seiner Entwicklung etwa dem Peter'schen Ei entspricbt. Verh. Anat. Ges. Erg. Heft des Anat. Anz. Bd. 37, S. 116-126. 1910. Flint, J. M. : Tbe Blood-vessels, Angiogenesis, Organogenesis, Reticulum, and Histology of the Adrenal. Jobns Hopkins Hospital Reports. Vol. 9. 1900. The Angiology* Angiogenesis, and Organogenesis of the Submaxillary Gland.

Amer. Journ. of Anat. Vol. 2, p. 417-444. 1903. The Development of the Lungs. Amer. Journ. of Anat. Vol. 6, p. 1-137. 1906. Fransen, J. W. P. : Le systeme vasculaire abdominal et pelvien des primates.

Petrus Camper. Deel 4. S. 215. 1907. Frederic, J. : Beitrag zur Anatomie und Entwicklungsgeschichte der Aste der aorta descendens beim Menschen. Schwalbes Morph. Arb. Bd. 7. 1897 auch Med. Diss. Strasburg 1898. Fuchs, H. : Zur Entwicklungsgeschichte des Wirbeltierauges. Anat. Hef te. Bd.

28. Heft 83. Wiesbaden 1905. Goppert, E. : Die Beurteilung der Arterienvarietaten der oberen Gliedmasse bei den Saugetieren und beim Menschen auf entwicklungsgesckichtlicher und vergleichend anatomischer Grundlage. 22 Fig. Ergebn. d. Anat. u. Ent wicklungsgesch. Bd. 14. 1905. Variabilitat im embryonalen Arteriensystem. Verhandlungen der anat. Ge sellschaft. 22. Versammlung. Erg.-Heft z. Anat. Anz. Bd. 32, S. 94.

1908. Ueber Riickbildung und Ersatz der arteria brachialis bei Echidna. Morph.

Jahrb. Bd. 33. 1909. Ueber die Entwicklung von Varietaten im Arteriensystem. Morph. Jahrb. Bd.

40. Heft 2 u. 3, S. 268^10. 1909. Grape, E. : Beitrage zur Entwicklung der Urniere und ihrer Gefasse beim Hiihn chen. Arch. f. mikr. Anat. Bd. 67. 1905. Grosser, O. : Zur Frage der segmentalen Gef assversorgung der Haut beim Menschen. Morph. Jahrb. Bd. 33, S. 555-561. 1905. Der Nerv des funften Visceralbogens beim Menschen. Anat. Anz. Bd. 37, S. 333-336. 1910 1 . Zur Kenntnis des ultimobranchialen Korpers beim Menschen. Anat. Anz.

Bd. 37, S. 337-342. 1910 2 . Hammar, J. A. : Ein beachtenswerter Fall von Halskiemenfistel. Ziegler's Beitr. z.

path. Anat. Bd. 36. 1904. Henneberg, B. : Verhalten der Umbilicalarterien bei den Embryonen von Ratte und Maus. Anat. Anz. Bd. IS, S. 321-324. 1900. Ueber die Anatomie und Physiologie der Umbilicalgefasse des Menschen. 34 Ber. Oberhess. Ges. f. Nat. u. Heilk. Med. Abt. Bd. 2, S. 11. 1901. Hepburn, D. : Halichoerus grypus, the Gray Seal. Journ. of Anat. and Phys. Vol.

30. N. S. Vol. 10, p. 498. 1896. ' Hill, E. C. : First Appearance of the Renal Artery. Johns Hopkins Hosp. Bui.

Vol. 16, p. 60. 1905. Hirsch, C. : Ueber die Entwicklung der Hornhautgefasse. Verh. d. Ges. deutsch.

Naturf . u. Arzte in Karlsbad 1902. Bd. 2. Heft 2. S. 382-383. (Erschien 1903). His, W. : Zur Bildungsgeschichte der Lungen beim menschlichen Embryo. Arch.

f. Anat. u. Phys. Anat. Abt. S. 89-106. 1887. Hoche, A. : Vergleichend-anatomisches iiber die Blutversorgung der Ruckenmark substanz. Zeitschr. f. Morph. u. Anthrop. Bd. 1, S. 241-258. 1899.


Hochstetter, F. : Ueber die Entwicklung der a. vertebralis beiru Kaninchen, nebst Beinerkungen iiber die Entstehung der ansa Vieussenii. Morph. Jahrb. Bd 16, S. 572-586. Taf. 21. -1890. Ueber den Ursprung der arteria subclavia der Vdgel. Morph. Jahrb. Bd. 16.

1890. Ueber die urspiiingliehe Hauptschlagader der binteren Gliedmasse cles Menschen und der Siiugetiere, nebst Benierknngen iiber die Entwicklung der aorta abdominalis. Morph. Jahrb. Bd. 16, S. 300-318. 1890. Entwicklungsgesckichte des Gefasssysterus. Ergebnisse d* Anat. und Entw. Bd. 1, S. 696. 1892. Die Entwicklung des Blutegefasssystems. Hertwig's Handbuch d. vergleicb. u. expt. Entwicklungslebre d. Wirbeltiere. 1906. (Erscbien 1903). Ueber den Ursprung der arteria caudalis beim Orajag und beim Kanincben. Anat. Hefte. Bd. 43. Heft 2. 1911. Hoffmann, M. : Zur vergleichenden Anatomie der Gehirn und Riickeninarkarterien der Vertebraten. Zeitsehr. f. Morph. u. Antbrop. Bd. 2, S. 247-322. Stuttgart 1900. Houssay, F. : La metamerie de l'endoderme et du systeme circulatoire primitif dans la region postbranchiale du corps des vertebres. C. R. Acad. Soc. Paris. T. 122. 1891. Hyrtl : Schlagadern des Unterschenkels. 1864.

Janosik, J. : Le developpenient des vaisseaux sanguins et des nerfs du membre anterieur cbez l'bomme et cbez quelques autres animaux. Arch, bohemes de Med. T. 4. 1891. Jeidell, H. : A Note on the Source and Character of the Early Blood-vessels of the Kidney. Anat. Record. Vol. 5. No. 2. 1911. Kadyi, H. : Ueber die Blutgefiisse des menschlichen Ruckenmarkes. Lemberg 1889. Kolster, R. : Studien liber die Niereno-ef asse. Zeitsehr. f . Morphol. u. Antbrop.

Bd. 4, S. 179. 1901. Krause, W. : Varietaten des Aortensystems und Varietaten der Korpervenen. Henle's Handbuch der syst. Anat. des Menschen. Bd. 3 : Gef asslehre. Braunschweig 1876. Leboucq : In der Disc, zu Zuekerkandl's Vortrag. Verb. d. anat. Gesellschaf t.

Erg.-Heft z. Anat. Anz. Bd. 8. 1893. Lehman, H. : On the Embryonic History of the Aortic Arches in Mammals. Anat. Anz. Bd. 26, S. 406^24. 1905. On the Embryonic Histoiy of the Aortic Arches in Mammals. Zool. Jahrb. Abt. f. Anat. Bd. 22, S. 387^34. 1905. Levi, G. : Morphologia delle arterie iliache. Arch. ital. di Anat. e di Embriologia.

Vol. 1. 1902. Lewis, F. T. : The Intra-Embryonic Blood-vessels of Rabbits from S l A to 13 days. Amer. Journ. of Anat. Vol. 3, p. 12-13. (Proc. Amer. Ass. Anat.) 1904. The Fifth and Sixth Aortic Arches and the Related Pharyngeal Pouches in the Rabbit and Pig. Anat. Anz. Bd. 28, p. 506. 1906. Locy, W. A. : The Fifth and Sixth Aortic Arches in Chick Embryos, with Comments on the Condition of the Same Vessels in other Vertebrates. Anat. Anz. Bd. 29. 1906. The Fifth and Sixth Aortic Arches in Birds and Mammals. Advance Print Proceedings of Seventh International Zoological Congress. Cambridge 1909. Luna, E. : Zur Morphogenese der unteren Zwerchfellarterien beim Menschen.

Arch. f. Anat. u. Phys. Anat. Abt. 1908. Macalister, A. : Multiple Renal Arteries. Journ. of Anat. and Phys. Vol. 17, p.

250. 1883. MacCallum, J. B. : Notes on the "Wormian Bodv of higher Mammals. Amer. Journ. of Anat. Vol. 1. 1902.

DEVELOPMENT OF THE VASCULAR SYSTEM. 705 Mall, F. P. : Development of tbe Human Coslom. Journ. of Morphol. Vol. 12, p.

395. 1897. Development Internal Mammary and Deep Epigastric Arteries in Man. Johns Hopkins Hospital Bulletin. 1898. On the Development of the Blood-vessels of the Brain in the Human Embryo.

Amer. Journ. of Anat. Vol. 1, p. 1. 1904. A Study of the Structural Unit of the Liver. Amer. Journ. of Anat. Vol. 5, p. 227. 1906. Manchot, C. : Die Hautarterien des mensehlichen Korpers. Leipzig. 1889. Manno, A. : Sur un cas interessant de "arteria saphena magna" chez l'homme.

Bibliogr. Anat. T. 14. Paris 1905. Martin, H. : Recherches anatomiques et embryologiques sur les arteres coronaires du coeur chez les vertebres. These Med. Paris 1894. Note sur le permier developpement des arteres coronaires cardiaques chez l'embryon du lapin. C. R. Soc. Biol. Ser. 9. T. 6. p. 83-85. 1894. Muller, E. : Beitrage zur Morphologie des Gef asssystems. I. Die Armarterien des Mensehen. Anat. Hefte. 1903. II. Die Armarterien der Saugetiere.

Anat. Hefte. 1904. III. Zur Kenntnis der Fliigelarterien der Pinguine.

Anat. Hefte. 1908. Parsons, F. G. : On the Arrangement of the Branches of the Mammalian Aortic Arch. Journ. of Anat. and Phys. Vol. 36, p. 389-399. London 1902. Pensa, A. : Studio sulla morph. e. s. sviluppo della a. intercost supr. Ric. f atte nel Lab. di Anat. umana normale della R. Univ. di. Roma etc. Vol. 2. Roma 1905. Osservazioni sulla moiphologia e sullo sviluppo della arteria pulmonalis nell'uomo. Boll, della Soc. Med. Chir. di Pavia, Communicazione fatta nella seduta del 8 Aprile 1910. Pavia 1910. Peter, K. : Mitteilungen zur Entwicklungsgeschichte der Eideehse. Arch, f . mikr.

Anat. Bd. 57. 1901. Ppeipfer, B.: Zur Kenntnis des histologisehen Baues und der Riickbildung der Nabelgefasse und des ductus Botalli. Virchow's Arch. Bd. 167, S. 210-231.

1902. Picchini, L. : Sviluppo dell' arteria pulmonare dell' aorta. II Morgagni. Vol. 33, p. 779-783. Napoli 1891. Popopf, D. : Die Dottersackgefasse des Huhnes. Wiesbaden. 1894. Popowsky: Ueberbleibsel der arteria saphena beim Mensehen. Anat. Anz. Bd. 17.

1893. Das Arteriensystem der unteren Extremitaten bei den Primaten. Anat. Anz.

Bd. 10.. Nr. 2, 3, 4. 1894. Contribution a la morphologie de 1'artere saphene chez l'homme. Bull, et Mem.

Soc. d'Anthrop. Paris 1904. Rabl, H. : Die erste Anlage der Arterien der vorderen Extremitaten bei den Vogeln.

Arch. f. mikr. Anat. Bd. 69. 1907. Rathke, H. : Ueber die Entwicklung der Arterien, welche bei den Saugetieren von den Bogen der aorta ausgehen. Mulleins Arch. f. Anat. u. Phys. S. 276.

1843. Ravn, E. : Uber die arteria omphalo-mesenterica der Ratten und Mause. Anat. Anz.

Bd. 10, S. 420-424. 1894. Rug : Anomalie der art. cruralis. Wiirzburger med. Zeitschr. Bd. 4, S. 344. 1863. Sabatier: Observations sur les transformations du systeme aortique dans la serie des vertebres. Ann. de Sc. Nat. Ser. 5. T. 19. 1874. Sakurai, T. : Zur Entwicklungsgeschichte der Lungenarterien. Anat. Anz. Bd.

25, S. 321. 1904.

Vol. II.— 45


Salensky, W. : Beitriige zur Entwicklungsgeschiehte der knorpeligen Gehorknochel chen bei Saugetieren. Morpk. Jahrb. Bd. 6. 1880. Salvi, G. : Arteriae superfieiales. Monitore Zoologico Italiano. T. 10. 1899. Schultze, 0. : Zur Entwicklungsgeschiehte des Gefasssystenis im Siiugetierauge.

Festschr. f. Kolliker. Leipzig 1892. Schwalbe, E. : Beitrag zur Kenntnis der Arterien- Varietaten des menschlichen Armes. Schwalbe's Morph. Arbeiten. Bd. 8. 1898. Semon, R. : Studieu iiber den Bauplan des UrogenitaLsy steins der Wirbeltiere.

Jenaische Zeitschr. f. Nat. Bd. 26. 1891. Soulie et Bonne : Contribution a Petude de l'appareil branchial et des arcs aortiques chez les mammiferes. Journ. de l'Anat. et de la Phys. Vol.

44. 1908. Sterzi, G. : Die Blutgef asse des Riickenmarks. Anat. Hefte. Abt. 1. Bd. 24, S. 3^364. Taf. 1-4. 1904. Tandler, J. : Zur Entwicklungsgeschiehte der Kopf arterien bei den Mammalia.

Morph. Jahrb. Bd. 30, S. 275-373. 1902. Zur Entwicklungsgeschiehte der menschlichen Daimarterien. Anat. Hefte.

Bel. 23, Heft 71. 1903. Ueber die Varientaten der arteria coeliaca und deren Entwicklung. Anat.

Hefte. Bd. 25. Heft 76. 1904. Ueber die Entwicklung des 5. Aortenbogens und der 5. Sehlundtasche beim Mensehen. Anat. Hefte. Bd. 38. Heft 115. S. 393. 1909. Thane, G. D. : Angiology, in Quain's Elements of Anatomy. Vol. 2. Part 2.

London 1892. Tonge, M. : Observations on the Development of the Semilunar Valves of the Aorta and Pulmonary Artery of the Heart of the Chick. Phil. Trans. 1869. Tonkoff, W. : Die Arterien der Intervertebralganglien und der Cerebrospinalnerven des Mensehen. Internat. Monat. f. Anat. u. Phys. Bd. 15. 1898. Die nervenbegleitenden Gefassnetze beim Embryo und die arteriae nutriciae nervorum beim Erwachsenen. Anat. Anz. Bd. 30. 1907. Vebsari, R. : Morfologia dei vasi sanguigni ai-teriosi dell' ochio dell' uomo e di altri mammiferi. Rieerche fatte nel Lab. d. Anat. umana norm, della R.

univ. di Roma. Vol. 8. Fase. 3-4. 1900. La morphogenese des vaisseaux sanguines de la retine humaine. Periodico del Lab. d. Anat. norm, della R. univ. di Roma. Vol. 10. Fase. 1. 1903. Ueber die Entwicklung der Blutgefasse des menschlichen Auges. Anat. Anz.

Bd. 35, Nr. 4. 1909. Vriese, B. de : Reeherehes sur revolution des vaisseaux sanguins des membres chez l'homme. Arch. Biol. T. IS, p. 665-730. 1902. Ueber die Entwicklung der Extremitiiten bei den Saugetieren. Verh. d.

Anat. Ges. 16. Vers. Halle a. S. Erg.-Heft z. Anat. Anz. Bd. 21. 1902. Sur la signification morphologique des arteres cerebrates. Arch. Biol. T. 21, p. 357-455. 1905. Reeherehes sur la morphologie de l'artere basilaire. Diss.-Inaug. Univ. Gand.

1905. Young, A. H. : On the Termination of the Mammalian Aorta, with Observations on the Homologies of the Pelvic Arteries. Stud. Anat. Dept. Owens College. Vol. 1, p. 209-225. 1891. Abnormalities of the Middle Sacral Artery and their Morphological Significance. Journ. of Anat and Phys. Vol. 31; and Studies in Anat. from Anat. Dept. Owens College. Vol. 2, p. 41. 1897. Observations on the Lumbar Arteries. Journ. of Anat. and Phys. Vol. 39, p. 295; and Studies in Anat. from Anat. Dept. University of Manchester.

Vol. 3. p. 129. 1906. Young, A. H., and Robinson, A.: The Posterior End of the Aorta. Journ. of Anat. and Phys. Vol. 32. 189S.

DEVELOPMENT OF THE VASCULAR SYSTEM. 707 Observations on the Development and Morphology of the Tail. Brit. Med. Joum. Vol. 2, p. 1384-1391. 1904. Young, A. H., and Thompson, P.: Abnormalities of the Renal Arteries, with Remarks on their Development and Morphology. Joum. of Anat. and Phys. Vol. 38. 1903. Zagorsky: Mem. de l'acad. des Seienc. de St. Petersbourg. Vol. 1, p. 326. 1809. Zimmekmaxx, W. : Ueber die Kiemenarterienbogen des Menschen. Verb. d. 10. internat. med. Kongr. Berlin. Bd. 2. Abt. 1. 1S90. Ueber emen zwischen Aorten und Pnlmonalbogen gelegenea Kiemenarterienbogen beim Kaninchen. Anat. Anz. Bd. 4, S. 720. 18S9. Zuckerkaxdl : Ueber die Entstehung der Vorderarnigefasse beim Kaninchen und bei der Katze. Verb. Anat. Ges. Erg.-Heft z. Anat. Anz. Bd. 8, S. 126-129. 1893. Zur Anatomie und Entwieklungsgesehichte der Arterien des Vorderarmes.

Anat. Hefte. Bd. 4, S. 1-98. 1894. Ebenda. Bd. 5. 1895. Zur Anatomie und Entwieklungsgesehichte der Arterien des Untersehenkels. Anat. Hefte. Bd. 5. 1895.

Histogenesis of the Arterial Wall.

Argaud, R.: Recherches sur la structure des arteres chez l'homme. 4 Taf. Toulouse 1903. Sur la .structure du canal arteriel chez le foetus du oeme mois. Toulouse 1905. Aschoff, A. : Beitrag zur Entwieklungsgesehichte der Arterien beim menschlichen Embryo. Schwalbe's Morph. Arb. Bd. 2. Heft 1. S. 1-35, auch Diss.

Strassburg 1892. Ueber Entwicklungs-, Waehstums- u. Alters- Vorgiinge a. d. Gefassen vom elast. u. musk. Typus. Jena 1909. Grunstein, N. : Ueber den Bau der grosseren menschlichen Arterien in verschiedenen Altersstufen. Arch. f. mikr. Anat. Bd. 47. 1896. Ueber den Bau der Wandung der Blutgefasse. Sitz.-Ber. d. Niederrh. Ges. f.

Naturw. u. Heilk. in Bonn. 10. Febr. 1S96. Morpurgo, B. : Ueber die Entwicklung der Arterienwand. Sitz.-Ber. d. kais. Akad.

d. Wiss. Wien 1885. Priebatsch, K. : Ueber die Histogenese der Aortenwand der Saugetiere mit be sonderer Beriieksichtigung der elastischen Fasern. Diss. Bern 1903. Roeder, H. : Die Histogenese des arteriellen Ganges, ein Beitrag zur Entwicklungs meehanik der Petalwege. Arch. f. Kinderheilk. Bd. 33, S. 149-161. 1902.


Anikiew. A.: Zur Frage iiber die Entwicklung der vena anonyma sinistra. Anat.

Anz. Bd. 34, S. 24. 1909. Bardeleben, K. vox: Die Hauptvene des Ai'mes, vena capitalis braehii. (Ueber die Entwicklung der Extremitatenvenen des Menschen.) Jenaische Zeitschr.

f. Naturw. Bd. 14. 1879. Benigxi, E. : Persistenza della vena ombilicale nelF adulto. Gazz. Ospidali. Anno 25. N. 85, p. 894-895. Bonnot and Seevers: On the Structure of a Human Embryo Eleven Millimetres in Length. Anat. Anz. Bd. 29, S. 452-459. 1906. Broman, I. : Ueber die existenz eines bisher unbekannten Kreislaufes im embi'yona len Magen. Anat. Anz. Bd. 23, S. 390-391. 1903. Rtrow: Beitrag zur Gefiisslehre des Fetus. Mfdlcr's Arch. S. 44. 1838. Davis. D. M. : Studies on the Chief Veins in Early Pig Embryos and the Origin of the Vena Cava Inferior. Amer. Joum. of Anat. Vol. 10, p. 161-472. 1910.


Dexter, F. : On the Vitelline Vein of the Cat. Anier. Journ. of Anat. Vol. 1, p.

261-267. 1902. Duval, M. : Apropos des veines ombilieales. Bull. Soe. Biol. Paris 1889. Gosset, A. : Contribution a l'etude du developpement de la veine cave inf erieure et des veines renales. Bull. Soe. Anat. Annee 73. Ser. 5. Vol. 12. Paris 1898. Grosser, 0., and Brezina, E. : Ueher die Entwieklung der Venen des Kopfes und des Halses bei Reptilien. Morph. Jahrb. Bd. 23. 1895. Zur Anatomie und Entwicklungsgeschichte des Gefasssystems der Chiropteren.

Anat. Hefte. Bd. 17. Heft 55. S. 131. 1901. Die Elemente des Kopfvenensystems der Wirbeltiere. Verb. d. Anat. Ges. Erg.-Heft z. Anat. Anz. Bd. 30, S. 179-192. 1907. Hochstetter, F. : Ueber die Bildung der hinteren Hohlvene bei den Saugetieren. Anat. Anz. Bd. 2. 1887. Ueber den Einfluss der Entwieklung der bleibenden Niere auf die Lage des Urnierenabsehnittes der hinteren Kardinalvenen. Anat. Anz. Bd. 3. 1888. Ueber die Entwieklung der Extremitatenvenen bei den Amnioten. Morph.

Jahrb. Bd. 17. 1891. Beitrage zur Entwicklungsgeschichte des Yenensystems der Amnioten. Sanger.

Morph. Jahrb. Bd. 20. 1893. Entwieklung des Venensystems der Wirbeltiere. Ergebn. d. Anat. u. Entwick lungsgesch. Bd. 3. 1893. Zur Entwieklung der venae spemiaticae. Anat. Hefte. Bd. 8 (Heft 27). 1898. Hoffmann, C. K. : Zur Entwicklungsgeschichte des Yenensystems bei den Selaehiern. Morph. Jahrb. Bd. 20. 1893. Kolliker, A. : Entwicklungsgeschichte des Mensehen und der hoheren Tiere. Leipzig 1879. Lewis, F. T. : The Development of the Vena Cava Inferior. Amer. Journ. of Anat. Vol. 1, No. 3, p. 229-244. May 1902. The Gross Anatomy of a 12 mm. Pig. Amer. Journ. of Anat. Vol. 2, p. 211 222. Mch. 1903. The Development of the Lymphatic System in Rabbits and the Development of the Veins in the Limbs of Rabbit Embryos. Amer. Journ. of Anat. Vol. 5, p. 113-120. Dec. 1905. On the Cervical Veins and Lymphatics in Four Human Embryos. Amer. Journ. of Anat. Vol. 9. Feb. 1909. Luschka, H. : Die Venen des menschliehen Halses. Denkschrift. d. Wiener Akademie. Bd. 20. 1862. Die Anatomie des Mensehen. Bd. 2, S. 341. Tubingen 1863. McClure, C. F. W. : A Contribution to the Anatomy and Development of the Venous System of Didelphys. Part 2 : Development. Amer. Journ. of Anat. Vol. 5, p. 163. 1906. Mall, F. P. : On the Development of the Blood-vessels of the Brain in the Human Embryo. Amer. Journ. of Anat. Vol. 4. Dec. 1904. On the Structural Unit of the Liver. Amer. Journ. of Anat. Vol 5. 1906. Marshall, J. : On the Development of the Great Anterior Veins in Man and Remnants of Fcetal Structures found in the Adult, a Comparative View of these Great Veins in the Different Mammalia, and an Analysis of their Occasional Peculiarities in the Human Subject. Phil. Trans. Roy. Soe. London 1850. Minot, C. S. : On the Veins of the Wolffian Bodies in the Pis:. Proc. Boston Soe.

Nat, Hist. Vol. 28 and Science N. S. Vol. 7. 1898.^ Parker, G. H. and Tozier, C. H. : The Thoracic Derivatives of the Postcardinal Veins in Swine. Bull. Mus. Comp. Zool. Harvard 1898.

DEVELOPMENT OF THE VASCULAR SYSTEM. 709 Pee, P. van: Note sur le developpement du systeme veineux du foie chez les em bryons de lapin. Journ. de l'anat. et de la phys. Vol. 35. Paris 1899. Piersol, G. A. : Human Anatomy. Philadelphia 1907. Popoff, D. : Die Dottersackgef asse des Huhnes. Wiesbaden 1894. Rex, H. : Beitrage zur Morphologie der Saugerleber. Morph. Jahrb. Bd. 14, S.

517-617. 188S. Salvi : Sopra lo sviluppo delle meningi cerebrali. Pisa 1897. Salzer, H. : Ueber die Entwieklung der Kopfvenen des Meerschweinchens. Morph.

Jahrb. Bd. 22. 1895. Schawlowski, I. : Zur Morphologie der Venen der oberen Extremitat und des Halses. Dokt.-Diss. St. Petersburg 1S91. (Compare Stieda, ref. in Ergebn. d. Anat. u. Entw. von Merkel u. Bonnet. Bd. 3, S. 380. 1894.) Smith, W. H. : On the Development of the Superficial Veins of the Body Wall in the Pig. Amer. Journ. of Anat. Vol. 9. July 1909. Soulie, A., et Bonne, C. : Reeherehes sur le developpement du systeme veineux chez la taupe. Journ. de l'anat. et de la phys. Vol. 41, No. 1, p. 1-39. 1905. Tandler, J. : Die Entwieklung der Lagebeziehung zwischen n. accessorius und v.

jugularis interna beim Menschen. Anat. Anz. Bd. 31, S. 473-480. 1907. Thompson, P. : A Note on the Development of the Septum Transversum and the Liver. Journ. of Anat. and Phys. Vol. 42, p. 170-175. 190S. Wertheimer, E. : Reeherehes sur la veine ombilicale. Journ. de l'anat. et de la phys. Vol. 22. Paris 1886. Williams, L. W. : The Somites of the Chick. Amer. Journ. of Anat. Vol. 11. No.

1, p. 55. 1910. Woodland, W. : A Suggestion concerning the Origin and Significance of the Renal-Portal System, with an Appendix relating to the Production of Subabdominal Veins. Proc. Zool. Soc. Lond., p. 887-901. 1906. Zumstein, J. : Zur Anatomie und Entwieklung des Venensystems des Menschen.

Anat. Hefte. Bd. 6. 1896. Zur Entwieklung des Venensvstems beim Meerschweinchen. Anat. Hefte. Bd.

8. 1897. Ueber die Entwieklung der vena cava inferior bei dem Maulwurfe und bei den Kanincheu. Anat. Hefte. Bd. 10. 189S.


Recent work on the development of the lymphatics has given us a new conception of the general morphology of the system as a whole. It has related the lymphatics to the vascular system and separated them from the system of tissue spaces. The study of human embryos x has sharpened this conception and made it possible to go a step farther, — namely, dividing the development of the system into two stages. The primary stage consists of a series of isolated lymph-sacs, which are clearly derived from the veins, and which become united into a system through two agencies, — (a)

1 Many of the facts concerning the development of the lymphatic vessels in human embryos have been obtained from the study of the Mall collection.


by the thoracic duct, which connects these sacs with each other, and (b) by the formation of a secondary opening into the veins at the jugular valves. The secondary stage involves the peripheral growth of lymphatic vessels which sprout from the endothelial lining of these sacs and spread out over the body. The invasion of the body is gradual, and in certain areas never takes place, as, for example, the central nervous system and the skeletal muscles. Since this new conception is not wholly accepted, — in fact, since most of the texts on anatomy and zoology describe the lymphatics as arising out of tissue spaces, — the evidence for the conception presented here will be given in detail as well as certain important general conclusions. 2 The first evidence of the formation of the lymphatic system is the development of symmetrical sacs in the neck, which have been called the jugular sacs. These are found first in a human embryo 10.5 mm. long (S.l.j., Fig. 488) as endothelial-lined sacs just lateral to the internal jugular veins (Sabin, 1909). In the same year Lewis (1909) described the jugular lymph-sacs in four human embryos, finding the beginning of the sac in an embryo 10 mm. long, in which it consisted of a single sac against the vein. In an embryo of 11.5 mm. he found four or five of such small sacs. His four stages are shown in excellent figures. He called attention to the fact (which is, I think, quite clear) that Ingalls (1908), in tracing the origin of the sac in an embryo 4.9 mm. long, was confusing veins and lymphatics. This jugular sac remains as the only sac until the embryo is 20 mm. long. The sac is formed in the following manner. Along the course of the jugular vein in early stages there is a series of branches which form a capillary plexus. Much of this capillary plexus disappears entirely, not being used to form the permanent branches. This destruction of capillaries is one of the fundamental factors in the evolution of 2 The fact that until very recently the weight of evidence rested on the side of the theory that the lymphatic system arose from tissue spaces will be shown in the following quotation from the last — that is, the 6th — edition of Kollikei-'s Geweblehre, 1902, page 681, " Ranvier glaubt daher, dass die Lymphgefasse vom Venensy stern nach der Peripherie in ahnlicher Weise durch Sprossung fortwachen, wie eine Druse mit verzweigtem Gangssysteme von einer Schleimlianti-ohre aus . . . Die Aufstellungen Ranvier's sind keineswegs sicher erwiesen und stehen im Gegentheile in Widersprueh mit den anderen gefundenen Tatsachen ; sie wurden jedoch hier angef uhrt, weil durch dieselben der Vorstellung von der ganzlichen Vershiedenheit von Bindegewebespalten und echten Lymphgefiitsen der scharfste Ausdruch gegeben wirt." Ranvier's comparison of the growth of the lymphatic system to the growth of a gland seems an unfortunate one, since the truer and more obvious comparison of the growth of lymphatic capillaries to blood capillaries, both invading by the same method, is thereby lost sight of. The second point brought out by Von Ebner, that, should the new theory prevail, it would lead to the sharpest possible separation of the lymphatics and the tissue spaces from the anatomical stand-point, is exactly what has happened.

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 711 the vascular system. In certain places, and first along the jugular vein, at the level of the primitive ulnar and cephalic veins, in embryos between 8 and 10 mm. long, some of the capillary plexus becomes cut off from the parent vein, and remains for a short time as a group of isolated endotlielial-lined spaces close to the vein. The extent of this zone, which probably varies considerably in different specimens, can be seen in Fig. 489, which is from a reconstruction of the jugular sac of the same embryo shown in Fig. 488. These isolated capillaries, the anlage of the lymphatic system, gradually dilate and coalesce to form symmetrica] endothelial-lined



Fig. 488. — Transverse section through the neck of a human embryo 10.5 mm. long, showing the symmetrical jugular lymph-sacs. (Mall's collection, No. 109.) X about 36. A., artery; N.S., nervus sympathicus; N.X., nervus vagus; Oe., oesophagus; P., pericardial cavity; S.l.j., saccus lymphaticus jugularis; T., trachea. The jugular veins are filled with blood and lie just medial to the lymph-sacs.

sacs, which subsequently rejoin the vein in such a way as to form a valve at the opening (Fig. 492). The time of the formation of the valve is in embryos between 10.5 and 12.5 mm. in length.

It is now necessary to prove that these jugular sacs are lymphatics, and. as this involves the use of the injection method on abundant material, it could not be done on human embryos. Conclusive proof that the jugular sacs are a part of the lymphatic system is readily obtained by injecting the lymphatics in the skin of the neck of other mammalian forms, as for example pig embryos, and proving that the lymphatic vessels empty into the sacs. This can be done in pig embryos from 18 to 20 mm. long. Below this stage the sacs could be identified in specimens in which the blood-vessels had been injected. After the position of the sacs had been determined, it was found that direct puncture of the sacs was the best method of obtaining extensive injections of the lymphatics which radiate out from them, thus indicating not only that the sacs are lymphatics but that they are important centres for the radiation of the lymphatic ducts.

That there are two large sacs in the neck of young sheep embryos, and that these sacs are lymphatics, was noted by Saxer (1896). Saxer. however, represents the theory, together with Gulland (1804), that the lymphatics come from tissue spaces, finding that the first lymph-vessels are in the subcutaneous tissue and are present in bovine embryos 25 mm. long. In 1000 Sala described the origin of the posterior lymph-sacs close to the veins in chick embryos. He

712 had, however, no conception of the significance of this discovery; if the earliest lymphatics are sacs close to the veins, the foundation is laid for the theory that the lymphatics grow from the veins to the periphery. Sala says that the

Fig. 489. — Reconstruction of the right jugular lymphatic sac, shown in solid black against the jugular vein, in a human embryo 10.5 mm. long. (Mall's collection, No. 109.) X about 14. G.N.V., Gasserian ganglion; S.v., sinus venosus; V.c, vena cephalica; Y.c.i., vena cava inferior; V.h., vena hepatica; V.j.i., vena jugularis interna; V.p., vena portse; V.p.c, vena cardinalis posterior;, vena subcardinalis; V.u.(p-), vena ulnaris (primitiva); V.u., vena umbilicalis; TF.6., Wolffian body.

posterior sacs arise from the veins and again that they are tissue spaces, two statements which mutually exclude each other. In addition he finds that the thoracic duct arises as solid cords of cells which secondarily become hollowed out into tubes and join the veins. If this be true, it can not, as Von Ebner

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 713 says in Kolliker's Geweblehre (Bd. 3. p. 682) in regard to these results of Sala's, " Wohl nicht bezweifelt werden, dass die Milchbrustgange beim Hiiehehen selbstandige Bildungen sind und nicbt aus den Blutgefassen hervorsprossen." But it is quite certain that in mammalian embryos the thoracic duct never arises as a solid column of cells. To return to the lymph-saes, their significance as the first lymphatics, together with the fact that the lymphatics grow from centre to the periphery, lays the foundation for the new theory as was brought out by myself in 1901 in the study of the system in pig embryos.

F. T. Lewis then showed, in 1906, that in rabbit embryos the jugular sacs are immediately preceded by a plexus of blood capillaries, so that they themselves are transformed capillaries. During the same year (1908) this method was confirmed in pig embryos by the method of injection by myself and in eat embryos by the method of wax plate reconstruction by Huntington and McClure together. 3 In pig embryos between 10 and 13 mm. long the entire plexus of capillaries external to the jugular vein can be injected from the vein, while in embryos 13 to 14 mm. long the plexus injects less and less from the vein until the sacs are formed. In one specimen the sac itself on one side received some of the ink which had been injected into the vein, showing conclusively that the sacs come from the capillary plexus. In pigs from 15.5 to 16 mm. long the sacs are never injected from the veins, and hence they are either entirely cut off, which condition lasts a short time, or the opening is guarded by a valve. Huntington and McClure (1910) traced this process by a complete series of wax models of the jugular region in cat embryos. The capillary plexus which is the anlage of the sacs, they called " veno-lymphatics." It may therefore be considered as proved that the jugular sacs are lymphatics and that they are transformed veins. The proof that they are the only lymphatics for a considerable time, until the embryo is 20 mm. long, that none of the tissue spaces, coelom, or the arachnoid spaces are a part of the lymphatic system, will be taken up later, in connection with the general consideration of the relation of the lymphatic system to tissue spaces.

The extension of the sac along the jugular vein may be by the addition of more of the capillary plexus, as is suggested by Figs. 490 and 491. These two figures are coronal sections from an embryo 11 mm. long. If they are superimposed, which can readily be done by matching the curve of the arm bud and the cephalic vein, it will be seen that the capillary plexus, the anlage of the lymph-sac, extends from the root of the primitive ulnar vein along the internal jugular vein into the neck. The full series shows that the plexus also extends a short distance into the arm bud along the primitive ulnar vein. The linear extent of the plexus is about 1.2 mm., an increase over the length of the preceding stage, which was 0.7 mm. The plexus is filled with blood, as if the secondary opening had not yet formed, and indeed, though the place of the valve is indicated in Fig. 490 by the projection of the sac into the 'Huntington and McClure in 1907 had advanced the view that the lymphatics came from clefts between the intima of the veins and the connective tissue, calling these clefts "extra-intirnal" anlages; but they retracted this theory, as far as the jugular sacs were concerned, during the following year (1908), and accepted the idea that the jugular sacs are venous in origin, though they think that the rest of the lymphatics are either "extra-intirnal" or of tissue space origin.

'14 angle between the internal jugular and cephalic veins, no break in the endothelium could be made out. This fact — that the endothelium shows no break in sections— is the only evidence on which rests the idea that the permanent opening is secondarily acquired. This contrasts sharply with the condition shown in Fig. 492 from an embryo 17 mm. long. The valve is first definitely open in an embryo 12.5 mm. long, as stated in the table on page 733, at which time the plexus has been transformed into a definite, long, empty sac of the type shown in Fig. 492. The valve is formed by a projection of the lymph-sac deep into the cleft between two veins, and it is so placed as only to be clearly evident in coronal sec

Arm bud >^VV^-'fe^lVi.l


"K v - u - l P }

?/££&': :W£ tfi.Trte B

A m

Fig. 490. — Frontal section through the arm bud of a human embryo 11 mm. long, to show the developing lymphatic sacs along the internal jugular vein. (Mall's collection, No. 353.) X about 37. S.l.j., saccus lymphaticus jugularis; V.c, vena cephalica; V.j.i., vena jugularis interna.

Fig. 491. — Frontal section through the arm bud of the same embryo as Fig. 489, to show the relation of the lymphatic-sac anlage to the primitive ulnar vein. X about 37. S.l.j., saccus lymphaticus jugularis; V.c, vena cephalica; V.j.i., vena jugularis interna; V.t.l., vena thoracica lateralis; V.u.(p.), vena ulnaris (primitiva).

tions. In transverse sections, as can be readily noted by comparison with Fig. 492, the valve is simply a tiny vessel between two larger veins; in sagittal sections it is even more difficult to locate.

There is no increase in the length of the sac in embryos between 12.5 and 17 mm. long, but from now on there is a rapid increase in size up to its maximum, which is reached in an embryo 30 mm. long, when the size is 5 mm. in length. This stage, which is an important landmark in several ways, is shown in a series of four figures, — 493, 494, 495, and 501. Fig. 493 is a reconstruction



of the primitive lymphatic system in an embryo 30 mm. Jong and shows that the stage which marks the maximum development of the jugular sac shows also all the other sacs and that they have been united into a complete system through the thoracic duct. Theperipheral system is also well under way, even much more than is shown in the figure, for the two stages of the lymphatic system — namely, the primitive central system of sacs and the peripheral system of ducts — overlap in their development. The position of the jugular sac can be seen by comparing Figs. 493 and 494. The level of the section shown in Fig. 494 corresponds with the line on Fig. 493. Both figures show the great size of the sac, it being by far

SJ finl

J'- ffin& 9 • 1

Fig. 492. — Frontal section through the arm bud of a human embryo 17 mm. long, to show the open valve of the jugular lymph-sac in relation to the veins. (Mall's collection, No. 296.) X about 26. S.l.j., saccus lymphaticus jugularis; V.c, vena cephalica; V.j.i., vena jugularis interna; Y.u.(p), vena ulnaris (primitiva).

the largest vascular structure in the neck. As shown in Fig. 493 it is now pierced by branches of three of the cervical nerves, namely the third, fourth, and fifth. These nerves help to orient the sac.

In the embryo 17 mm. long there was a slight extension of the jugular sac into the arm bud. This extension is now much larger, making a definite subclavian sac (S.l.s.) along the primitive ulnar vein (V.u.p.). 4 The jugular sac in this stage shows two other important points, — namely, its relation to peripheral lymphatics, and an ex 4 This origin of the subclavian sac in human embryos as an extension of the jugular sac is interesting in connection with F. T. Lewis's (1906) discovery, that in rabbits the subclavian sac arises independently from the veins.


tensive bridging of its dorsal border, which is the process by which the sac is transformed into a chain of lymph-nodes. These two processes are closely related in function. In Fig. 493 one enormous superficial lymphatic vessel {V.l.s.), which arises from the lateral surface of the sac, extends out to the skin, and spreads out into a plexus of large capillaries in the subcutaneous layer. One of the smallest of these superficial lymphatics is shown on the left side of Fig. 494.

This group of vessels is the first set of lymphatics to reach the skin. This has been abundantly proved in pig embryos by many injections into the skin (Fig. 507). In pig embryos this set of vessels reaches the skin in the neck at about 18 mm.; in human embryos about 20 mm. long. At this stage no injection of any layers of the skin in any other place except the neck has ever entered lymphatics. The great size of these early lymphatic vessels to the skin is in some sense represented in the adult by the greater size of the vessels of the deep subcutaneous plexus in contrast with the superficial plexus, and calls to mind the size of the subcutaneous lymph-sacs of the amphibia.

In Fig. 493 are seen lymphatics extending over the skin of the head as a superficial plexus (V.l.s.) and deep lymphatics (V.l.p.) extending from the subclavian sac along the course of the primitive ulnar vein into the arm bud.

The bridging of the jugular sac along its dorsal border is shown in Fig. 495. The level of this section is also indicated on Fig. 493. This process of bridging or the cutting of the lumen of the sac by bands of connective tissue begins early, being first noted in an embryo 14 mm. long. It is a process by which the sac, originating from a plexus of blood-capillaries, is reconverted into a capillary plexus this time lymphatic in character. This lymphatic plexus is far more extensive than the preliminary bloodcapillary plexus, as may be seen by comparing the early sac of Fig. 489 with the one in Fig. 493, from which the lymphatic plexus is formed, or by comparing the length of the blood-capillary plexus along the vein, 0.3 to 0.7 mm., with the length of the sac, 5 mm.

To complete the account of the jugular sacs as far as they have been studied — that is, up to the stage when the fetus measures 80 mm.— the sac becomes more and more encroached upon by the connective-tissue bridges, until it is transformed into a plexus of lymphatic capillaries, out of which chains of lymph-glands are evolved.

In Fig. 493, beside the jugular-subclavian sac, there are three other sacs, the retroperitoneal, the posterior, and the cisterna chyli. None of these sacs nor any anlage of them has been made out in embryos under 20 mm. in length. The retroperitoneal sac and cisterna chyli are present in a human embryo 23 mm. long, while all three are present in one 24 mm. long.



Fig. 493. — Profile reconstruction of tie primitive lymphatic system in a human embryo 30 mm. long. (Mall's collection, No. 86.) X about 5.8 C.c, cisterna chyli; L.g., lymphoglandula; N.III., N.IV., and if. V., nervi cervicalis; S.l.jug., saccus lynphaticus jugularis; S.l.mes., saccus lymphaticus retroperitonalis; S.l.p., saccus lymphaticus posterior; £./.s., saccus lymphaticus subclavius; Y.c, vena cephalica; V.e.i., vena cava inferior; Y.f., vena femoralis; V.j.i., vena jugularis interna; Y.l.p., vasa lymphatica profunda; Y.l.s., vasa lymphatica superficialia; Y r., vena renalis; Y.s., vena sciatica; V.u.(p.), vena ulnaris (primitiva I.

'18 The retroperitoneal sac was discovered as a part of the lymphatic system by F. T. Lewis (1901-02 and 1906). Baetjer (1908) found in carefully tracing its history in embryonic and fetal pigs, that it is preceded in embryo pigs 17 to IS mm. long by a plexus of capillaries in the root of the mesentery, which drained into the large anastomosing vein at the hilum of the two Wolffian bodies. These capillaries are readily injected from the vein, as is seen in Fig. 496. The same figure shows the large renal anastomosing vein between the Wolffian bodies ventral to the aorta. It also shows well the mass of connective tissue between the vein and the mesentery in which the retroperitoneal sac develops. The plexus retains its connection with the veins until the embryo is 20 mm. long, as is shown in Fig. 497. Other sections of the same series showed more of the ink within the plexus, but this section was chosen to show the connection with

Fig. 494. — Frontal section through the jugular lymph-sacs in a human embryo of 30 mm. (Mall's collection, No. 86.) X about 9. The level of the section is shown on the reconstruction of Fig. 493. The section shows the complete lymph-sac on the right side and the valve on the left. S.l.j., saccus lymphaticus jugularis; VS., v. innominata; V.j.i., v. jugularis interna; V.l.s., vasa lymphatica superficialis; Oe. , oesophagus; T., trachea.

the vein. From .this time on, the plexus is readily transformed into a sac, as shown in Fig. 498 in an embryo 23 mm. long, in which the sac is entirely separate from the vein. The sac joins the cisterna chyli, through which it can drain into the thoracic duct and the veins, in embryos 27 mm. long (Fig. 499). These four figures are the best representation that we have of the proof of the transformation of a venous plexus into a lymphatic sac.

In human embryos the stage corresponding to Fig. 496, in which there is a plexus of veins ventral to the renal anastomosis, has been identified in an embryo 20 mm. long, while at 23 mm. there is a definite retroperitoneal sac and a cisterna chyli. The retroperitoneal sac lies in the root of the mesentery adjacent to the great masses of the suprarenal bodies and the sympathetic nervous system (S and G.s., Fig 500). It extends from a point



opposite the fourth lumbar vertebra anteriorly, to the point where the superior mesenteric artery enters the mesentery. The position of the retroperitoneal sac is also shown in Fig. 493 and its relation to the renal vein in Fig. 501. which corresponds with the line so marked on Fig. 493. The sac has never been found as large in human embryos as it is in the pig. In a fetus 80 mm. long it is represented by a long chain of lymph-glands or a plexus of lymph-vessels which form the anlage of the glands ventral to the aorta. It has recently been shown by Heuer (1909) that injections of the retroperitoneal sac enable one to follow the progression of vessels from this sac out into the mesentery along

V5* iT ^ ;*'

t . " . > , Fig. 495. Frontal section through the jugular lymph-sac of the same embryo, at the level shown In Fig. 493, to show the bridging of the sac which is the anlage of the first lymph-gland. X about 19. S.l.j., saccus lymphaticus jugularis.

Fig. 496. Transverse section through the renal anastomosis of the subcardinal veins of an embryo pig 18 mm. long. (After Baetjer.) X about 43. A., aorta; G.A., genital anlage; M.C., retroperitoneal capillaries; Mes., mesentery; R.A., renal anastomosis; W.B., Wolffian body.

the superior mesenteric artery. Within the mesentery is formed a secondary great lymphatic plexus, the anlage of the lymphoglandulas mesentericae (Lg.m.), as shown in Fig. 502. From the mesenteric vessels lymphatics gradually invade the intestinal wall.

The posterior lymph-sac has as yet been identified only in pig and in human embryos among mammals. It has, however, been worked out in chick embryos by Sala (1900), where it is a true lymph-heart with muscle in its wall, as in the amphibia. Sala's work has already been referred to; it is the most recent work based on the theory, also brought out by Gulland (1S94) and Saxer (1896),

-At »9

720 that the lymphatics arise from tissue spaces, unless one includes the work of Huntington and McClure who hold a modified form of this theory (1910). Sala found that the posterior lymph-hearts begin at the middle of the seventh day in connection with the lateral branches of the first five coccygeal veins. He says that corresponding to these veins there are excavations in the mesenchyme which soon enter into communication with the lateral branches ("E sono rappresentati da spazi scavati nel mesenchima che sta laterahnenti ai miotomi caudali, a livello delle prime cinque vene coccygee," p. 292) ; and in fact one would say that these fissures are simply dilatations of the veins themselves (" Si direbbe anzi che esse non sono che semplici dilatazioni, ramificazioni delle stesse vene," p. 269). These two statements, of course, contradict each other, for spaces can not be both fissures in the mesenchyme and dilatations of the veins. Then he

Fig. 497. — Transverse section through the renal anastomosis of the subcardinal veins in an embryo pig, 20 mm. long. (After Baetjer.) X about 43. In this section the venous channels in the root of the mesentery are beginning to show definite evidences of fusion and sac formation, though they are still connected with the veins, as is shown in the figure. A., aorta; G.A., genital anlage; M.C., retroperitoneal capillaries; Mes., mesentery; R.A., renal anastomosis; W.B., Wolffian body.

describes these fissures as becoming more abundant and confluent. By opening up communications with each other they form a sac or lymph-heart in the mesenchyme. This sac, he says, is lined with flattened mesenchyme cells, which, if it were so, would, according to our stand-point, exclude it from being a vein. He found muscle in the wall of the hearts on the ninth day, and was able to inject the heart directly by the second half of the tenth day. Sala's description of the origin of the posterior lymph-hearts in the chick is, nevertheless, so clear and graphic, and corresponds so closely with the method of origin of the lymphsacs in mammals, that one easily suspects that the two processes are the same, — that the sacs arise from the veins in both cases. The fact that Sala uses the description as evidence of the old conception, of the lymphatic system as coming from tissue spaces, does not necessarily confuse the picture. Mierzejewski (1909), working on the chick, has confirmed Sala's description, and evidently is of

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 721 the opinion that both Sala and he find the sac arising from the veins, for he states that "An den Enden der lateralen Aeste der ersten fiinf Coccygealvenen bilden sieh kleine, blasenartige Ausbuchtungen, die sich bestandig vergrossern und am siebenten Bebriitungstage eine Reihe von segmental nacheinanderfolgenden, mit den Vene in Verbindung stehenden Spalten im embryonalen Bindegewebe bilden. Diese Anlagen des spateren Lymphherzens nehmen im Verlauf der Entwicklung an Grosse zu und nahern sich einander immer mehr und mehr, so das sie schlieszlich miteinander an den Stellen, wo sie beriihren verschmelzen." He also states that he agrees with Sala, except that the process begins a little earlier than Sala described, — namely, in the middle of the 6th rather than the beginning of the 7th day. Thus it seems a fair conclusion that the weight of evidence from the study of the posterior hearts in the chick is on the side of their venous origin.

  • s j

In a human embryo 20 mm. long there is a plexus of capillaries along the v. ischiadica which forms the anlage of the posterior sac.

Fia. 498. — Transverse section through the rena anastomoses in an embryo pig 23 mm. long. (After Baetjer.) X about 53. This is the first appearance of a definite sac in the exact location of the venous plexus in the earlier stages. It will be noticed that the irregular margins suggest the fusion of many small vessels. At this stage no connection can be traced between the sac and either the lymphatic system or the veins. A., aorta; M.S., retroperitoneal sac; R.A., renal anastomosis; W.B., Wolffian body.

The saccus posterior or ischiadicus is first found in an embryo 24 mm. long, and is well shown in Fig. 493 in the embryo 30 mm. long. Here it is a long narrow sac — seen also on one side in Fig. 501 — which extends along the external surface of the v. ischiadica primitiva (S.l.p.), from the posterior end of the cisterna chyli to the bifurcation of the v. femoralis and the v. ischiadica. The sac reaches an apparent maximum in fetuses 80 mm. long. It is shown to great advantage in sagittal section in a fetus 80 mm. long in Fig. 504. The posterior sac is now clearly a pelvic structure, being transformed into a chain of lymphoglandulae iliacae.

Vol. II. -46

722 The question of the origin of the cisterna chyli and the thoracic duct has proved a difficult problem because the region is hard to eject.

Recent studies on pig embryos throw some light on the thoracic duct. In a pig embryo 23 mm. long (measured fresh along the mesencephalosacral line; compare Chap. VIII) the left jugular sac was filled with ink through the superficial lymphatics. The needle was then withdrawn and pressure applied to the head. By a fortunate chance most of the ink ran over into the thoracic duct while very little ran into the veins. Usually the ink passes readily through

Fig. 499. — Transverse section through the early cisterna chyli and retroperitoneal sac in an embryo pig 3 cm. long. (After Baetjer.) X about 40. The section shows the connection of the cisterna chyli and the retroperitoneal sac, by large channels along the lateral margins of the aorta. A., aorta; K., kidney; I., intestine; M.S., retroperitoneal sac; R.C., cisterna chyli; P.C.V., postcardinal vein.

the valve into veins. In this specimen it can be shown that the jugular sac anlage of the thoracic duct has three connections with the jugular sac, and passes as a plexus of lymphatic vessels toward the median line between the sympathetic nerve and the common jugular vein (these relations can be made out in Fig. 48S). The two ducts, the thoracic duct and the right lymphatic duct, extend in the loose connective tissue dorsal to the oesophagus about to the level of the arch of the aorta. At this stage there are no lymphatics corresponding to the eisterma chyli, but there are especially abundant median anastomoses of the posterior cardinal or azygos and hemiazygos veins dorsal to the aorta opposite the adrenal anlages.



In the next stage — namely, in an embryo 25 mm. long — the jugular sac anlages are more extensive and now symmetrical, for the right duct has turned ventralward toward the root of the lung while the left, or thoracic duct, remains near the aorta. A second change of importance has taken place, — namely, the separation of a new lymph-sac from the veins, the cisterna chyli, which exactly replaces the previous plexus of veins. An abundant plexus of lymphatic vessels encircles the aorta from this cisterna chyli, so that one can not in this region speak of a right and left duct, but rather of an aortic plexus. This fact is interesting in comparison with Pensa's (1908 to 1909) figures showing the comparative morphology of the duct in various forms, for he shows that in a number of forms the lower part of the thoracic duet is an abundant plexus of lymphatic vessels. By the time the pig is 27 mm. long the relations of the three prevertebral

Fig. 500. — A composite diagram made by superimposing the sections showing the relations of the retroperitoneal sac and cisterna chyli to the veins, in a human embryo measuring 27 mm. (Mall's collection, No. 382.) X about 8. A.m.s., a. mesenterica superior; C.c, cisterna chyli; G.s., ganglia sympathies; S.l.m., saccus lymphaticus retroperitonealis; S., suprarenal body; Y.a., v. azygos; v. c.i., vena cavainferior.

lymphatic anlages are established, the right lymphatic jugular sac anlage, making the right duct, has reached the root of the lung, while the left jugular sac anlage has anastomosed with the cisterna chyli anlage along the aorta, making the thoracic duct. All the specimens studied show some isolated endothelial-lined spaces which cannot be traced to connect with the thoracic duct in serial sections. The existence of these isolated spaces was pointed out by Lewis (1906), and, since he could trace a few of them to join veins, he suggested that there might be multiple venous anlages of the lymphatic vessels analogous to the lymph-sacs. Since these spaces or isolated islands are found along other veins as well as the azygos veins, they will be discussed under the general considerations (p. 737). The other recent work on the thoracic duct is by McClure (1908), in which he agrees with Lewis that the thoracic duct arises by multiple anlages from the veins. This view he retracted in 1910 in favor of the extra-inthnal theory; but, since he did not retract the evidence in his first paper but only the interpretation of the observations, it must be stated that he confused veins and lymph

724 aties, calling certain veins lymphatic anlages, when it is easy to demonstrate that these same veins persist as veins after the thoracic duct is formed.

In human embryos the observations on the thoracic duct are still scanty. In the embryo shown in Fig. 488, which is 10.5 mm. long, there is on the left side a small vessel extending from the



Fig. 501. — Frontal section through the retroperitoneal sac of the human embryo, at the level indicated on Fig. 493. X about 40. A., aorta; G.s., ganglia sympathica; K., kidney; S.l.m., saccus lymphaticus retroperitonealis; S.l.p., saccus lymphaticus posterior; V.c.i., vena cava inferior; V.r., vena renalis.

sac toward the median line. In an embryo 16 mm. long there are symmetrical jugular sac anlages of the thoracic and right lymphatic ducts, which, however, do not reach the zone dorsal to the oesophagus. The first appearance of the cisterna chyli is in an embryo 23 mm. long, as shown in Fig. 500. Here it is a definite sac opposite the third and fourth lumbar vertebrae, at



the point where the vena cava curves ventralward and where it anastomoses with the azygos vein. By the time the embryo is 30 mm. long, as shown in Fig. 493, the thoracic dnct is complete. The lower or cisterna chyli portion is much simpler than in the pig embryos, in fact there is a right and a left vessel, and the right duct crosses behind the aorta in the thorax to join the left. Thus the human embryos illustrate the double origin of the duct from the jugular sacs on the one hand and from the cisterna chyli, a true lymph-sac, on the other.

Fio. 502. — Transverse section through the abdominal cavity of a human embryo 80 mm. long. (Mall's collection, No. 172.) X about 8. A.m. a., arteria mesenterica superior; C.c, cisterna chyli at its lower border; Lg.m., lymphoglandulse mesenteries ; S.l.m., saccus lymphaticus retroperitonealis.

The question of the peripheral growth of lymphatics is one which really lies at the root of the new conception of the origin of the lymphatic system from the veins. In studying the peripheral lymphatics it was found that they converged toward or radiated out from certain centres. These centres proved to be the lymphatic sacs. The lymphatic sacs become united into a system by means of the thoracic duct, and connected with the veins by the development of valved openings. The sacs and thoracic duct may be termed a primary system, which is shown on Fig. 493. The secondary or peripheral lymphatics, according to our view, grow out from the primary system. In tracing the peripheral

726 lymphatics we may refer again to Fig. 493, which shows not only the primary system of sacs complete but the beginning of the peripheral vessels. The earliest peripheral lymphatics that have been made out in a human embryo are those from the jugular sac to the skin of the neck in an embryo 20 mm. long. These vessels are shown (V.l.s.) in the embryo 30 mm. long in Fig. 493, and again as the large deep vessel behind the ear and pointing toward the shoulder in Fig. 505. At the stage of 30 mm., beside the superficial vessels for the back of the head, there are deep lymphatic vessels (V.l.p.) extending along the subclavian vein. The transformation of the jugular subclavian sac into a chain

Fig. 503. — Sagittal section of a human embryo measuring 50 mm., showing the posterior lymph-sao within the pelvis and its extension along the femoral vein. (Mall's collection, No. 96.) X about 8. F., femur; La-, lymphoglandula (femoralis); O.s., os sacrum; S.l.p., saccus lymphaticus posterior with lymphgland in the border; V.s., vena sciatica; V.I.V., vetrebra lumbalis V.

of lymph-glands makes the primary group of glands for these two sets of vessels. In Fig. 493 is shown a small gland (Lg.) on the course of the plexus of lymph-ducts at the edge of the subclavian sac. This marks the beginning of secondary glands, that is those that form on the course of lymph-ducts. From the posterior sac two sets of peripheral vessels are extending, one along the v. femoralis, shown as V.l.s., while the second set, which follows the V. ischiadica to the hip, is not shown. The vessels along the v. ischiadica have, however, reached the skin and spread out over hip and back at this stage, as can be seen for the pig in Fig. 507.

Recently I have had the privilege of studying a remarkable specimen of a lymphatic distention in a human embryo. The embryo, which is 5.5 cm. long, was injected through the umbilical artery bv Professor Max Broedel while the heart was still beat



ing. It was then placed in formalin and left there for about a year. Dr. H. M. Evans then began to study the vascular injection in the skin vessels, and while working on it put the embryo into freshly made-up 50 per cent, alcohol. To his amazement, there appeared a wonderful injection of air in the skin, which proved to be a complete injection of the superficial lymphatic system. The irregular lymphatic plexus shown in silvery lines was of great beauty. Two tracings were made of the specimen under direct sunlight, with the aid of a camera lucida, one a side

Fig. 504. — Transverse section through the pelvis of a human embryo 80 mm. long, to show the posterior lymph-sacs. (Mall's collection, No. 172.) X about 9. S., bladder; Lg., lymphoglandula; if., rectum;, saccus lymphaticus posterior.

view shown in Fig. 505, the other a dorsal view, Fig. 506. The injection gradually disappeared, but for a few days could be restored by the use of fresh alcohol.

The specimen is of the same stage as the largest pig embryo figured in my article (1904) on the superficial lymphatics, and the two specimens make an interesting comparison. It is the stage of the single primary lymphatic plexus, and only in one area, namely in front of the ear, was there a double plexus, deep and superficial.

728 Notwithstanding the great irregularity of the plexus, a quite definite pattern is to be made out. The vessels all drain into two areas, as shown on the side view. First the vessels from the head, neck, arm, and thorax run toward the jugular-subclavian sac, and secondly those from the leg, hip, and abdominal wall run toward the groin to the posterior sac. These points of drainage

Fig. 505.

Fig. 506

Fig. 505. — Distention of the lymphatic vessels with air of a human fetus 5.5 cm. long, drawn by means of a camera lucida. (Mall's collection, No. 448.) X about 2. The drawing had to be completed without the object, a-b, area without lymphatics.

Fig. 506. — Distention of the lymphatic vessels with air in the same fetus as in Fig. 505.

are marked on the surface by the few large definite trunks that radiate toward them. The vessels which drain toward the neck and axilla — namely, the trunk behind the ear and the pectoral trunks below the arm — have valves. The vessels on the abdominal wall pointing toward the groin are large and irregular, but are as yet without definite valves. Valves are wanting in all the rest of the vessels. It is striking how completely the entire lymphatic plexus anastomoses, so that theoretically one can inject the

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 729 entire lymphatic system through any one vessel whatever. The natural flow of lymph toward the sacs, however, is indicated by the size of the main trunks.

The extent of the injection is most interesting. There is a small area on the head in the mid-line (between the letters a and b in Fig. 505) which never showed any lymphatics. It is the same area, only less extensive, that could never be injected in the pigs of this stage, and thus is probably entirely free from lymphatics. The fingers, toes, the palms of the hands, and soles of the feet likewise had no injection whatever.

The pattern of the vessels over the head shows a number of interesting points. Over the face the mesh is much finer than over the scalp. The eyelids show a few vessels, the ear none. Behind the ear is shown a large deep trunk, seen in both figures, which drains the back of the head and neck and undoubtedly enters the jugular sac or gland. This trunk is to be compared with the vessel marked V.l.s. in Fig. 493 and in Fig. 494. There is probably also a deep, large channel in front of the ear, for the vessels of the face and chin converge there, but the double plexus of capillaries was so dense there that none could be made out. On the back of the head, long parallel vessels drain toward the jugular trunk on either side, while in the centre the plexus is fine-meshed toward the top of the head and coarse-meshed toward the neck. The vertebra prominens is marked by being rather free from lymphatics, and the same is true of the bony prominences at the elbow and ankle.

The arm shows a fine-meshed plexus; the vessels reach the cleft between the fingers in each case. The pattern of the forearm is made by long parallel vessels running lengthwise, while in the upper arm the vessels run around toward the axilla.

The plexus over the ventral surface of the body is fine-meshed, and there is a complete anastomosis across the mid-line ; over the back the plexus is coarse-meshed. A few large vessels over the chest and back drain toward the axilla; a similar set converge to the groin. The latter do not yet show valves. The especial characteristic of the back region is that the mid-line is bridged by long, rather slender parallel vessels. This is more marked in the lower third.

On the foot the vessels reach the clefts between the toes just as on the hand they reach the clefts between the fingers. The malleoli are quite free from vessels. On the leg the vessels run obliquely rather than lengthwise on the lateral aspect, while on the thigh the plexus points toward the groin.

The double injection of blood-vessels with Prussian blue and of the lymphatic capillaries with air enabled one to see their relative positions with great clearness. The large main blood-vessels

730 of the skin were deeper than the lymphatics, while the entire system of smaller arteries and the blood-vascular capillary plexus lay superficial to the much larger lymphatic plexus.

The development of the peripheral lymphatics out from the sacs to the ultimate capillaries has been worked out in the skin of the pig, of the bird, (Mierzejewski, 1909), and of bovine embryos (Polinski, 1910). In the skin (Sabin, 1904) a great number of injections have brought out the fact that the vessels spread out from two great centres, the neck and the groin, so that the vessels gradually extend from lymphatic to non-lymphatic areas. Fig. 507 will

Fig. 507. — The lymphatic vessels of the skin of an embryo pig 4.3 cm. long. X about 2H- The injected vessels form the primary subcutaneous plexus and represent a complete injection except in the area dorsal to the ear, — that is to say, the uninjected areas have not yet received lymphatics.

serve to show the spreading out of the lymphatics in the primary subcutaneous plexus. The group in the neck is growing out from the jugular sac, the group over the hind leg is extending from the posterior sac. Both of these injections are complete or nearly so, showing that there is a large non-lymphatic area at this stage. Later a secondary, finer-meshed, and more superficial -plexus develops. From the retroperitoneal sac, the peripheral spread of the lymphatics to the ultimate lacteals of the intestine has been worked out by Heuer (1909) in the pig. Fig. 50S, taken from this paper, shows the entrance of the groups of lymphatics from the mesentery into the intestinal wall, and the primary submucosal plexus not yet complete. Later a finer-meshed plexus forms in the mucosa, and from this plexus the lacteals grow into the villi.



This point of the gradual progression of the lymphatics within an organ out to the ultimate eapilliaries, which is one of the strongest proofs of the growth of lymphatics from the veins to the periphery, rather than from the periphery to the veins, is also very convincingly shown by H. M\ Evans. Fig.

Fio. 508. — Loop of the small intestine of an embryo pig 100 mm. long, to show the growth of the lymphatics into the intestine, and the formation of a primary submucosal plexus out of a series of lymphatic loops. (After Heuer.) 509 is taken from this paper, which describes a case of sarcoma of the intestine in which there is a growth of new lymphatics out into the tumor mass. It will be noticed that the growth is from the mucosal or capillary plexus, that at the edge of the tumor the new vessels are like normal lacteals, while within the

Fig. 509. — A piece of adult intestine showing a sarcomatous nodule which is being invaded by growing lymphatic capillaries from the mucosal, plexus. The transition from the normal lacteal; to the new vessels is to be noted in passing from right to left. The large submucosal ducts are seen in the shaded area. (After Evans.)

nodule there is an over-development of lymphatics in the form of an advancing plexus. It has also been pointed out to me by Dr. Evans that in the adult intestine the valves of the lymphatics occur at the base of the capillary bed, that is in the submucosal duets, and that they mark the place of transition between the duct


and the capillary. Thus the mucosal plexus and the lacteals are the ultimate capillaries. This agrees with the general theory of Ranvier, that in the lymphatic system vessels without valves have the structure of capillaries. The discovery that the lymphatic vessels invade each organ, that the invasion can be demonstrated by injections of successive stages as soon as the line of growth or point of entrance is known, together with the fact that the valves develop at the base of the capillary bed, gives us the key by which the relations of the lymphatic system within each organ can be worked out from the primary distributing plexus of ducts to the ultimate capillaries. It may be well to note here that the lymphatics as they enter an organ are always capillaries, that is the growing zone is always the capillary bed.

To sum up, the peripheral spread of lymphatics thus far observed in human embryos, from the jugular-subclavian sac, two sets of vessels extend, one to the skin of the head, neck, and shoulder, the other as deep lymphatics to the arm. From the posterior sac two sets of vessels develop, one along the v. ischiadica to the skin of the hip and back, a second set along the v. femoralis to the leg (Fig. 493). The retroperitoneal sac sends vessels into the mesentery (shown as Lg.m. in Fig. 502). On the course of these vessels a mass of lymph-glands develops and vessels extend out from these glands to the intestine. The cisterna chyli drains both the retroperitoneal and the posterior sacs (Fig. 493). The progression of the lymphatics in the Mall collection is summed up in the following table.

In regard to the development of lymph-glands the series of human embryos serves to establish an interesting general relation and to illustrate certain phases in the development of an individual gland. In general, the first stage in the development of an embryonic lymph-gland is the formation of a plexus of lymph-ducts. This was one of the earliest points established and goes back to the time of Breschet (1836). The first lymphgland to appear is through the transformation of the jugular lymph-sac into a plexus of lymph-vessels. This bridging of the sac is shown in Fig. 495, and is simply a reduction of the sac into a plexus of lymph-vessels lined with endothelium, with bridges of connective tissue between, in which the mesenchyme is slightly denser than in the surrounding tissue. All the primary lymphsacs are thus transformed into a plexus of lymph-vessels. In the jugular sac the transformation extends over a series of embryos and fetuses from 14 to 80 mm. long. The bridging of the retroperitoneal sac is illustrated in Figs. 501 and 502, in the posterior sac in Figs. 503 and 504. The cisterna chyli becomes bridged only along its borders 5 (Fig. 502). Primary groups of glands may be

5 In this connection it is interesting to note, that in the amphibia the pulsating lymph-hearts have exactly the same relation to the peripheral vessels as the sacs in mammals have to the corresponding vessels. Thus the sacs and primary lymph-glands represent the amphibian lymph-hearts.



Length of tinembryo in QC Ci Ci Cn ifc CO to to tc to *- H- k


1— t 1— 1 h- • 1— » k^

O O O © o> O if* CO c O OS M ^1 OS en Cn 4* to i-» O CO CO

Cn in


t— ' to


to ^ if* if* co H* CO CO K P co Xo. in the -J to CO tO CO M * * » Cn CO O 05 / to to to •*) 00 to If* CO o O C5 o CO to CO cn O if* if* ~4 Cn O O co co co CO collection.

H H 1-3


  • J


2 a 2 a ffl - - -_ •_

  • i p

3 P 5- < to <D a. "I CO CO p M 7 » f» ? g p l-l p

co p

gj p p p $ 2 gittal gittal nsvers .gittal nsven c* o p

3 3 CO CO <! < CD CD CO 00 3 g 5- < p CO a. ^ CO 3 CD < CO •-J 00 B CO CD i.



r* P_ o p E.

o s P Z X 7: 2- * •< B CD CO CD CO p 00 <: CD •-» m S- o a = 5 co co CD



c co

O P o 1

rr *• co ^ ^ C.T

^ c^ CO /

7^ \L * * ^ X

to P (-> *

1— '


1— » i— * to o r X Z. N

B? f* 1- H. M CO


a co bi


Cn in o E

— X ^j • 2.,_ 3 cn en ! S. <* b

1— * Qi

bo 2 CO ^

  • -• to


[> M •=> CO CO ^ .- & p p o <3 p a co cf 3 S ° C P §3 P CO O 00 c 5' 00 CO t-l p -1 CO CO «-( p _,. P e CO o CO TO f CC cP g -r "^ co tr 1 3 CB E p 3-^ P CO -J.

p o B f o p CO CO i-a . _P v; M P n 3 ^ a

with few ly turning int ny follicles, e bridges th hain of lym TO CO CO p o 1 c* p

< CO >-j •< CO B 2 ~3 EL p-B. p CD a -I - to 2.

CO ~ P- S cr o << CD X ^ CO >-J ' P VJ ELP » go 3 00 CO k

<; CO << 00 B p_ < CD v; 00 E — VI oo CO r CD TO 00 p o p" p CD kj CD _ 2.— K 47 EL™ c B s p ' == 2, 5 -j <0 2. 3^ S'g."' g s.

O &. CO CD ef ' - X O CD P o p tj o 5" 3 •a 3* P 5" 2

— c mph-n o lymp rougho ph-nod c" to M 3 CD oS.

<! p_ < CO JO

« p r+ CD

5 a -r-ct

= b' TO p" CO CD P

5T X p cn O

(TO P -! <7 2 c b* 2 A P

B £

5 E= DO


o 13 P iq <*• Ml

B CD p 2. ?b £.to CO 3 • p

CO CD 2 p

O TO •s CO

Ei b' TO CO

3- V< < — ' - CO 

o -

£ O CO a- CO m

< 00 ? S" p ^ E.

= ' 5" 3 P_ O X CD 3

c* CO "° c* 5 3

o' c "2. cT X CD ^_ CO g.cT 2.

O X CD g < s' CO § OE

o' 3

2. o'


c cn


2. <; CO • 3 5" o r* CO

CO <r* O P

o <! 2.

5' CO ^'O -60 TO S • pTO P*^ co


3 % CO





s <

1 •5. p"

e* CD P" CD o

c CD

<! p_ o .

p ^.

— c

p a,

P p

p p

< CD Sto 2 c




TO i

- p to o o •a 3 3 CD


t fou t fou mage roun m"co 00 co 3 S 00

P ><



• c* en




CO B ~ a. • • <r* P

3 "" o g


c" 1

co B.


<< p

O —





5" 5°

<^* 2.






3 CO

^ ?°

I P o a.


1 c

co O

CD 3 _

CO oo 3 1


gore's p g ro 3 c hd K

  • d

"1 -1 CO CD CO

DC CI X 1 > m g ,_, co CD CO CO CO


CO CO co

— g . B P o to • S- S

3 P P

= CO ct- c+ e*


- P- D 3

2 "~ << o o

^ 3 3 — —

3 — 5 S ^

T =

p- B p


i J. "3 ~

— P — 1 — O 7 7

S 3 OB a as


CO o o

go a, &



co to *- CO to g X -J x b x >-* x to

It* h3

to x S. ^p X __ coi^" bg ' -o

o a.




2. X 2 * r* 2 £ 2 p" CD C^ • CD in 5' B.

3 TO

<: p L.13

3. -! p 5'

o p TO

s cc W o t— I S5 o H B C?3


CO <J p »1 w CO 3 r o B ^ OQ S c* — H a ^ P5 H CO


O TO "3

3" CD

P P a. p_ a. p' 3


H H f > i— i O GO

W o p H s* 03 ° ?3 K! O > 2! a H a cc O •=1 W tr 1 tr* O o f s o H o

p 3 CO <S CD

P P a.

p CD

P" a a. o *i CO O < CD P



defined as those which develop from the lymph- sacs. They are the jugular-subcl avian chain, the retroperitoneal or pre-aortic chain, and the chain of lymphoglandnlse iliaea?. All the lymph drains through these chains. 6 Secondary lymph-glands develop around plexuses of the peripheral vessels, and two of these are shown in Fig. 493, one in the arm, and the second in the leg. The secondary group from the retroperitoneal sac is shown in Fig. 502 as the mesenteric group of glands (Lg.m.). The ultimate lymph-glands which develop at the base of the final capillary bed as the lymphfollicles of the intestine were not found in the series ; they are probably the last to develop. This is in harmony with the findings of Anton (1901) in connection with the lymph-glands of the Eus

Fig. 510. — Reconstruction of an axillary ymph-node anlage from a human embryo 70 mm. crovn-rump, showing the primary plexus of lymphatic capillaries. (After Kling.) X about 41.6. 6, the bands of connective tissue between the lymphatics; these are shown as darker than the lymphatics; d, dorsal; I, lymphatic vessels; Is, blind end of a sprouting lymphatic capillary.

tachian tube and the middle ear. He found no glands there in the fetus, while during the first two years of life there was a gradual development of lymphocytes which subsequently formed definite follicles.

In regard to the development of an individual node, the important stages can be well illustrated in human embryos. 7 The

" This idea finds a very interesting' confirmation in the work of Jolly (1910) on the lymph-giands of birds, for he finds that the first lymph-glands, the jugular and ischiatic groups, are centrally placed, and thus support Ranvier s theory of the growth of lymphatics from centre to periphery.

7 The development of lymph-nodes has been followed in a number of recent papers by Saxer (1896), Gulland (1894), Kollmann (1900), Kling (1904), Sabin (1905 and 1909), and Jolly (1910), of which Kling and Sabin refer to human embryos.



first step in the formation of lymph-glands is a plexus of lymphatic capillaries, and this is true whether the gland forms out of one of the primary sacs or along the course of peripheral lymphatic vessels. This first stage of a lymph-gland is illustrated in section in Fig. 495 for the jugular lymph-gland in an embryo measuring 30 mm. long. The character of the lymphatic plexus is also well shown in Fig. 510, after Kling, from a reconstruction of the subclavian or axillary group of a fetus somewhat larger, measuring 70 mm. In the primary stage the lymph-node is wholly lymphatic in structure, — i.e., it consists of a plexus of lymphatic capillaries with undifferentiated connective-tissue septa.


Fig. 511. — Reconstruction of a lymph-node from a human embryo 270 mm. long, showing the peripheral and central sinuses. (After Kling.) X about 41.6. b, connective-tissue bands; sin.m., sinus marginalia;, intermediar sinus; vas.a., vas afferens; vas.e., vas efferens.

The second step of the development of the node is the heaping up of lymphocytes or wandering cells in the connective-tissue septa, forming follicles. These masses of cells are shown in Kling 's first model as the darker masses labelled b, some of which are oval, while more are irregular in shape. In our series the first definite follicles are found in a fetus 50 mm. long. The follicles are associated with a plexus of blood-capillaries. All the recent investigators note these blood-capillaries in the connective-tissue septa. Thus, in the second stage a lymph-gland contains two elements, a lymphatic element, or the plexus of lymphatic capillaries, and

736 a vascular element, consisting of blood-capillaries surrounded by lymphocytes in the meshes of the connective tissue, making the follicles. The follicles are well shown in Figs. 503 and 504. The first two stages, while they can be sharply separated in a series of early embryos, in later embryos develop side by side.

The third stage in the development of a gland is the formation of the sinus out of the plexus of lymphatic capillaries. That the sinus is a capillary plexus, as dense as the blood-vessels in cavernous tissue, is shown most beautifully for the adult in the injections of lymph-glands in Teichmann's Atlas. The reorganization of the node, the development of the peripheral and central sinuses, together with the great increase in the follicle, is well

9 iM

•C. b. c.

  • 9


•P. b. cf.


  • r

if f « 



>in. fit

Fig. 512. — Hemal node from the neck of an embryo pig 245 mm. long. B.v., blood-vessel at hilum; C.b.c, central blood-capillary; P.b.s., peripheral blood-sinus.

shown in Fig. 511, after Kling. It shows a model of the lymphatic part of a lymph-gland from a fetus 270 mm. long. The very great size of some of the vessels of the marginal sinus is to be noted. In the development of the various nodes the greatest possible variations occur in the proportion of the lymphatic element or follicle. The sinus formation is also shown in Fig. 504. The sinus differs from the primary lymphatic plexus in the extreme thinness of the connective-tissue septa. It remains to be shown whether it differs also in the nature of its endothelium, — that is, whether the sinus, which begins as a dense plexus of closed lymphatic capillaries in fetal stages, is a closed system in the adult or not.

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 737 In the alimentary canal there are certain special lymph-glands — namely the tonsils, solitary follicles, and Peyer's patches — that develop in the capillary bed close under the epithelium. In connection with these nodes there has been considerable confusion in regard to their development. This confusion was the more easy as long as it was thought that the thymus, derived as it is from epithelium, was lymphoid in character. Stohr (1S91 and 1898) and Kollmann (1900) have pointed out that the lymph-nodes in the mucosa of the alimentary canal are mesodermal in origin, as is all the rest of the lymphatic system, rather than ectodermal.

Hemal glands have not been found in human embryos. In pig embryos they appear only in late stages, the first and simplest type, shown in Fig. 512, being found in the neck of a pig 245 mm. long. Here the gland consists of a single follicle around a plexus of blood-capillaries and surrounded by a sinus of bloodvessels. It is thus possible to define the follicle as a collection of lymphocytes around a blood-capillary plexus. The follicle is surrounded by a sinus which may be made of a plexus of lymphatic capillaries forming a lymph-sinus, or by a plexus of blood-capillaries making a blood sinus. The lymphatic sinus is found in the lymph-glands, the blood-sinus is found in the hemal node and in the spleen. A group of lymph-follicles makes a lymph-gland; a group of blood follicles makes a hemal node. In shape the follicle is primarily round, but where the lymphocytes extend along the course of the artery the follicle becomes elongated into the ellipsoids of the spleen or the cords of the lymph and hemal glands. Meyer (1908) showed, by hundreds of injections of haemolymph or better hemal nodes, that they are not connected with the lymphatic system nor are they intercallated in the course of the veins.

General Considerations.

It is now necessary to bring out certain general considerations which follow from the recent studies on the lymphatic system.

The relation of the lymphatic system to tissue spaces has been one of the greatest questions in connection with the system since it was first vaguely suggested by Aselli (1622) about three centuries ago, and clearly formulated by Lieberkiihn (1760) in connection with the discovery of the central lacteals of the villi and the supposed opening of these lacteals into the connective tissue. The histoiy and bearing of this great question were best brought out by His (1863). The general conception of the morphology of the lymphatic system has passed through a series of phases. The early experiments of Nuek (1691), of injecting air into the arteries and noting its return through the lymphatics, led to the theory of the connection of the finest arteries and lymphatics. These hypothetical connections may be grouped together under one general term by which they were known, namely vasa serosa. The vasa serosa are associated with a variety of names, notably Boerhaave, Haller, and Bichat (1818). It was the theory of Haller that the vasa serosa were so small that the red blood-corpuscles could not pass through them, and hence only the fluid of the blood ran over into the lymphatics. As the methods of injecting the fluids were perfected it was noted that only in exceptional cases did fluid forced into the arteries enter the lymphatics. But these observations did not have as much weight in overthrowing the theory of vasa serosa as the development of the ideas of Schwann (1S39) and especially of Virchow (1863). From Schwann's observations on the capillaries in the tadpole's tail, he suggested, in connection with his discovery of cells in the animal body, that capillaries were a network of anastomosing cells making canals all over the body. Virchow overthrew the idea of the vasa serosa, as will be seen in the following quotation from the English translation of his Cellular Pathology (p. 76): "Amongst these different species of connective tissue, the most Vol. II.— 47


important for our present pathological views are, generally speaking, those in which a reticular arrangement of cells exists, or, in other words, in which they anastomose with one another. Wherever, namely, such anastomoses take place, wherever one cell is connected with another, it may with some degree of certainty be demonstrated that these anastomoses constitute a peculiar system of tubes or canals which must be classed with the great canalicular system of the body, and which particularly, forming as they do a supplement to the blood and lymphatic vessels, must be regarded as a new acquisition to our knowledge, and as in some sort filling up a vacancy left by the old vasa serosa, which do not exist." Thus Virchow substituted the idea of hollow connective-tissue cells to connect arteries and lymphatics for the vasa serosa. The methods of injection, however, led to sharper and sharper conceptions of the lymphatic capillary, and made, as His (1S63) says, the obscure lymphatic roots more and more of a myth. The beautiful injections of Teiehmann, together with his own work, led His (1861) to formulate the opinion that " Die ersten wurzeln des Systems durehweg der eigenen, isolierbaren Wand entbehren, es sind Kanale in das Bindegewebe der Cutis, der Schleimhaute usw. eingraben.* The next great step was the discovery that capillaries are lined with endothelium, one of the most important discoveries in histology. This dates back to the work of Hoyer in 1865. The names of Kolliker, Teiehmann, His, Hoyer, Ludwig, and von Recklinghausen are to be associated with the development of the conception of a lymphatic capillary as an endothelial lined structure, either in the form of a network or as blind ends like the lacteals. The introduction of silver nitrate injections by Hoyer (1865), His (1863), and von Recklinghausen gave a method by which the limits of the endothelium of the lymphatics were more sharply determined; but the silver-nitrate pictures led von Recklinghausen astray, as we believe, to a conception of lymph radicles or tissue spaces as a part of the lymphatic system. The stomata and stigmata by which the lymphatic vessels were thought to connect with the lymph radicles have not been confirmed, and are more and more clearly seen to be mechanical defects of the silver-nitrate method. The question now presents itself in two phases, first the relation of the lymphatics to the tissue spaces in general, and secondly to certain special tissue spaces like the piarachnoid. To the first important question, the theory that the lymphatics come from the veins has a perfectly clear and satisfactory answer, and may be considered to have settled a difficulty which has faced anatomists for three hundred years. The lymphatic sacs and capillaries have exactly the same relation to the tissue spaces as have the blood-capillaries. Both are foreign structures that grow into or invade the mesenchyme. Tissue spaces are no more a part of the lymphatic system than they are of the blood-vascular system. Thus, fluid within the veins should be called blood-serum, the fluid in the tissue spaces might be termed plasma, while the term lymph should be reserved for fluid within the lymphatics. The use of three distinct expressions as indicating three distinct elements would be a decided advantage. Von Recklinghausen's silver pictures show two different systems, the lymphatic vessels and the tissue spaces or lymph radicles. Melzer (1896 and 1911) has brought out well the physiological meaning of these distinctions.

In embryos before the formation of lymphatics, the mesenchyme varies greatly in different places, that is, it is considerably differentiated. In certain special constant places the meshes of the mesenchyme are very large, — for example, around the central nervous system. These spaces around the nervous system have especial significance in connection with the lymphatics, for these mesenchyme spaces, the anlage of the piarachnoid spaces, extend along the peripheral nerves in young embryos and have been confused with lymphatics.

Sections of human and other mammalian embryos will show spaces along the growing nerves, contracted at the origin of the nerves but widely expanded at the growing tips. These spaces may be termed perineural spaces.

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 739 In studying a long series of pig embryos injected into the piarachnoid space, it is found that often the injection mass runs out into the perineural spaces, thus outlining the peripheral nerves. Such injections do not enter true lymphatics, thus showing the independence of these two systems. In a study of the arachnoid made by the injection method in the Anatomical Laboratory of the Johns Hopkins University by L. L. Reford. and as yet unpublished, it has been shown that the thinning out of the mesenchyme around the central nervous system is not haphazard, but that injections of the same stage give the same pattern, and that the form of the arachnoid space changes as the brain develops. That is to say, the arachnoid space has as definite a form as the coelom and it never connects with the lymphatics. Moreover, no injections of lymphatics run over into the arachnoid or perineural spaces, showing that the great arachnoid and perineural space system is not a part of the lymphatic system.

Through the work of Budge (1880, 1887) there developed a theory that the coelom had a genetic relation or developed in common with the lymphatic system. He injected the extra-embryonal coelom in chick embryos, and found that the fluid passed out into the area vasculosa in forms simulating vessels and thought that this formed a primitive lymphatic system.

The finding that the lymphatic system arises from the veins, and that the tissue spaces and all the serous cavities of the body therefore stand in the same fundamental relation to the lymphatic system as they do to the blood-vascular system, marks a definite advance in our conception of the general morphology of the body, and is perhaps the most valuable result of the recent studies on the lymphatic system. This is as true for cavities like the various bursas and chambers in the eye as for the piarachnoid, the ccelom, the pleural and pericardial cavities.

Closely associated with the question of the relation of the lymphatics to tissue spaces are two points — namely, the question of growth of the lymphatic capillaries and the time-honored question of open and closed lymphatics — which are the most interesting of all problems associated with the structure of the lymphatic system.

The question of the growth of lymphatics is the crucial point in connection with the new theory. That the lymphatic capillaries and blood-capillaries grow by the same method was suggested by Kblliker as early as 1846 in a study of the living tadpole's tail. The matter could not be on a firm basis until after the important discovery that blood-vessels were lined by endothelial cells. The idea of the growth of blood-capillaries by sprouting had its earliest beginnings in the work of Schwann (1839), and involves a long series of observations in which the most telling are those on the living amphibian larva. That the growth of the lymphatic capillary, like that of the blood-capillary, is from the sprouting of their endothelial lining ceils was first discovered by Langer in 1S6S in a study of the tadpole's tail.

These observations, long unnoticed, were rediscovered by Ranvier (1S95-1S97) in a series of studies on amphibian and mammalian embryos. Ranvier saw that with this method of growth it was impossible to think of lymphatics starting as dilated tissue spaces and growing toward the centre. MacCallum (1902) was the next to call attention to this method of growth, and he added the observation, that, in watching the injection of these growing capillaries under the microscope, there were no lymph radicles connecting the lymphatics with the tissue spaces as seen in silver-nitrate specimens, but that the lymphatics had a complete wall and ruptured explosively under too great pressure, and were hence anatomically closed vessels. The lymph radicles are tissue spaces. Bartels (1909) has repeated the injections of MacCallum, and obtains the same figures of the long sprouts of endothelium ; he suggests, however, the theoretical objection to the theory growth of lymphatics by sprouting that this involves an idea of growth against the pressure of

740 the fluid contained (pp. 46-47). This theoretical difficulty is not a real one, since it is possible to watch the lymphatics grow in the living form. The final proof that lymphatics grow by sprouting of their endothelium has been given by E. R. Clark (1909 and 1911), who watched the growth of a given lymphatic vessel in the same tadpole's tail under the high powers of the microscope for long periods of time. He describes, that, from the sides and ends of the growing vessels, long processes of protoplasm push out into new territory; these processes now advance, now, bend far out of their course to pick up some stray blood-corpuscles, and now retract entirely like long slender pseudopodia. Moreover, by subjecting the tadpole to lower temperatures, the activity of the endothelium can be checked, while under the stimulus of heat numerous tiny threads of protoplasm are pushed out, only a few of which grow into permanent lymphatics. Moreover, he has added an important discovery on the nature of the endothelium. He finds that the growing tip consists of a hyaline membrane, in which the nuclear areas, that is nuclei hidden by granular protoplasm, divide and move up and down the wall

Fig. 513. — View of the blood-vessels and lymphatics (solid black) of a tadpole's tail (Hyla Pickeringii 10 mm. long). Fixation in Zenker's fluid. (After Clark, 1911.) The oblique lines represent the myotomes, and the numbers indicate the corresponding vessels of Fig. 514.

and out into the growing tips, even passing one another, so that the growing » tip is unquestionably a synecytium. This clears up one difficulty long associated with the idea of growth by sprouting — namely, whether endothelial strands were individual cells which subsequently became hollowed out. The tiniest vessels are hollow tubes of protoplasm. This discovery enlarges our conceptions of endothelium, especially in connection with Mollier's (1911) beautiful specimens showing the endothelium of the splenic veins in the form of a reticular protoplasmic syncytium.

It cannot be said that there is agreement among the recent workers on lymphatics. This "disagreement, we think, rests on a fact noted by His as far back as 1863, that " Eine nicht injizierte Lymphwurzelrohre zu erkennen, beinahe unmoglich ist." The lymphatic capillaries in early mammalian embryos seen in serial sections are conspicuously large in contrast with the blood-capillaries, but



they are at the same time extremely irregular and the largest vessels are often connected by the tiniest threads of endothelium. In 1906 F. T. Lewis showed in rabbit embryos certain small isolated spaces arranged hi bead-like rows along the primitive veins extending out from the regions of the primitive sacs. These vessels are probably lymphatics; they are lined by true endothelium, are empty, and are larger than blood-capillaries. Their interpretation is given in Clark's figures 513 and 514. The study of these spaces, in their relation to the lymphatic system, resolves itself into an analysis of the limits of error of different methods. The three methods employed have been the study of the living by Clark, the study of serial sections of iminjected embryos -by Lewis, Huntington, and MeClure, and of injected embryos by myself. That the study of serial sections yields valuable results is unquestioned; the observations on the general distribution of the lymphatics in human ernbryos were made on such material. But in determining the essential point — namely, the method of growth of the lymphatic tip, whether by the addition of connective-tissue spaces, or by numerous venous anlages, or by

Fig. 514. — The same area as in Fig. 513. (After Clark, 1911.) Reconstruction from serial sections 10 m thick, stained with hsematoxylin and Van Gieson's mixture (acid fuchsin and picric acid), made by the use of an oil-immersion lens (Zeiss obj. 2 mm. and ocular 6).

the growth of its own endothelium — the method of the interpretation of sections fails, because the point at issue lies within the limits of error of the method.

The relative value of the method of iminjected and injected sections was in part tested in the skin of the embryo pig (Sabin, 190S), but the whole question has been much more conclusively tested by Clark (1911) in the tadpole's tail. His figures, two of which are copied, show the essential points. He first studied the entire lymphatic system as it can be seen in the living tadpole's tail. There is no question but that all of the lymphatics, to the last endothelial cell and protoplasmic sprout, can be seen. Moreover, the entire system so seen can be injected and nearly as much can be made out in a total specimen in alcohol. Such a specimen was drawn (Fig. 513). The tadpole was then sectioned and exactly the same area reconstructed. It was found, in the first place, that the amount which could be reconstructed varied greatly with the intensity of the stain.


With weak stains, like eosin or congo red, comparatively little could be seen, but an intense fuchsin stain gave the maximum advantage. By using a high dry lens (Zeiss 4 mm.) in specimens stained in fuchsin, both blood-capillaries and lymphatics split up into isolated islands or Lewis anlages. With the oil-immersion lens (Zeiss 2 mm.) more of the vascular and lymphatic systems could be reconstructed, but the isolated vessels were more distal, but still numerous, as seen in Fig. 514. This figure represents the maximum amount of blood-vessels and lymphatics that can be reconstructed in the tadpole's tail under the favorable conditions of knowing the extent of the lymphatics in the exact area and of intense protoplasmic stains. Clark regards this, and I think properly, as a crucial test of the relative value of methods.

The work of Huntington and McClure (1908-1910), presented for the most part in joint publications, advances a different idea in regard to the lymphatic system. Their position is a complicated one, for they hold that the lymphatic system arises by three different methods : first, that the jugular lymph-sacs are venous in origin; secondly, that some of the peripheral lymphatics are clefts between the endothelium of the veins and the surrounding mesenchyme, their socalled extra-intimal space theory; and thirdly, that some of the peripheral lymphatics arise as tissue spaces. In regard to the first point, they originally thought that the jugular lymph-sacs were extra-intimal (1906-08), but abandoned that idea in 1908. The extra-intimal theory is not a serious obstacle in connection with the lymphatic problem. Huntington and McClure find in specimens which have been fixed in Zenker's fluid, that the endothelium of the veins shrinks away from the surrounding tissue. This phenomenon they find more common in veins which are degenerating. In studying through a series of human embryos which show a great variation in the amount of maceration and in quality of fixation, we have found that such spaces vary according to the fixation. Moreover, in studying human embryos, it becomes clear that there are certain constant areas of unusually loose mesenchyme and these areas are the first to show the effects of maceration. In the living tadpole no extra-intimal spaces are to be seen, while in fixed specimens they are present. Thus the extra-intimal space is open to the charge of artefact, and, in order to be taken seriously, must at least first be shown to occur with all of the best fixatives. In connection with the lymphatics it can be shown that the growing lymphatic does not always follow the vein. Here it should be emphasized that the blood-capillaries lie in perfectly definite and constant areas, so that whether lymphatics actually replace them or not is readily tested. The lymphatic sacs do replace the veins; some at least of the peripheral lymphatics which McClure (1910) claims are extra-intimal do not replace veins, but exist beside them. Huntington and McClure show a tendency to abandon the extra-intimal theory in favor of the old theory of the connectivetissue origin of all the lymphatics save the jugular lymph-sacs. The theory that lymphatics grow by the addition of tissue spaces rests on the observations of sections. The appearances in sections remain open to a variety of interpretations, in contrast with the simplicity and sharpness of the appearances in the living form. Therefore for the theory that the lymph-vessels develop otherwise than by the sprouting of the endothelium of preceding vessels we have no sufficient proof.


Anton, W: Studien iiber das Verhalten des lymphatischen Gewebes in der Tuba Eustachii und in der Paukenhole bein Fetus, beim Neugeborenen und beim Kinde. Zeitschr. f. Heilk. Bd. 22. 1901.

Baetjer, W. A. : On the origin of the Mesenteric Sac and Thoracic Duct in the Embryo Pig. Amer. Jour, of Anat. Vol. 8. 1908.

DEVELOPMENT OF THE LYMPHATIC SYSTEM. 743 Bartels, P. : Das Lymphgef asssystern, in Bardeleben's Handbuch der Anatomie des Meusehen. 1909. Bichat : Anatomie generale. Paris, 1801. Breschet: Le systenie lymphatique. p. 185. 1836. Budge: Ueber ein Kanalsystem im Mesoderm vou Hiihnerenibryonen. Arcb. f.

Anat. u. Pbys. Anat. Abt. 1880. Untersuebungen iiber die Entwicklung des Lympbsystems beim Hiihner embryo. Ebenda. 1S87. Chievitz: Zur Anatomie einiger Lympbdriisen im erwaebsenen imd fetalen Zustande. Arcb. f. Anat. u. Pbys. Anat. Abt. 1881. He describes the mesenteric glands in tbe buman embryo. Clark, E. R. : Observations on Living, Growing Lymphatics in the Tail of the Frog Larva. Anat. Record. Vol. 3. 1909. On the Inadequacy of the Method of Reconstruction in Studying Developing Lymphatics. Anat. Record. Vol. 5. 1911. Conil, C. E. J.: Contribution a l'etude du developpement des ganglions lympha tiques. These de Bordeaux. 1890. Engel: Ueber den Bau und die Entwicklung der Lymphdrusen. Prager Viertel jahrsschrift. Bd. 2. 1850. Evans, H. M. : On the Occurrence of Newly-formed Lymphatic Vessels in Malignant Growths. Johns Hopkins Bulletin. Vol. 19. 1908. Gulland: The Development of Lymphatic Glands. Journ. of Path, and Bact.

Vol. 2. 1894. Gives an excellent review of the literature. Rep. from the Labor, of the R. College of Physicians. Edinburgh. Vol. 5. 1896. Heuer, G. J. : The Development of the Lymphatics in the Small Intestine of the Pig. Amer. Journ. of Anat. Vol. 9. 1909. His, W. : Ueber die Wurzeln der Lymphgef asse in den Hauten des Korpers und iiber die Theorien der Lymphbildung. Zeitschr. f . wiss. Zool. Bd. 12. 1863. Ueber das Epitbel der Lymphgefasswurzeln und iiber die von Reckling hausen'schen Saftkanalchen. Zeitsch. f. wiss. Zool. Bd. 13. 1863 2 . Hoyer: Ein Beitrag zur Histologie bindegewebiger Gebilde. Arch. f. Anat. u.

Pbys. u. wissenschaftl. Medizin. 1865. Huntington : The Genetic Interpretation of the Development of the Mammalian Lymphatic System. Anat. Record. Vol. 2. 1908. The Phylogenetic Relations of the Lymphatic and Blood-vascular Systems in Vertebrates. Anat. Record. Vol. 4, p. 1. 1910. The Genetic Principles of the Development of the Systemic Lymphatic Vessels in tbe Mammalian Embryo. Anat. Record. Vol. 4, p. 399. 1910. Huntington and McClure: The Development of the Main Lymph-channels of the Cat in their Relation to the Venous System. Anat. Record. Vol. 1.

1906-1908. The Anatomy and Development of the Jugular Lymph-sacs in the Domestic Cat. Anat. Record, yol. 2. 1908. The Anatomy and Development of the Jugular Lymph-sacs in the Domestic Cat. Amer. Jour, of Anat. Vol. 10. 1910. Ingals, N. W. : A Contribution to the Embryology of the Liver and Vascular System in Man. Anat. Record. Vol. 2. 1908. Jolly, J. : Recherches sur les ganglions lymphatiques des oiseaux. Arch. d'Anat.

Microsc. T. 11. 1910. Gives a complete bibliography of Ranvieris works. Kling: Studien iiber die Entwicklung der Lymphdrusen beim Menschen. Upsala Lakareforenings Forhanslhiger. 1903 and Arch. f. mikr. Anat. u. Ent wicklnngsgesch. Bd. 63. 1904. Knower, H. McE. : The Origin and Development of the Anterior and Lymphhearts and Subcutaneous Lymph-sacs in the Frog. Anat. Record. Vol. 2.



Kolliker: Hist ologisc he Studien an Batrachierlarven. Zeitsehr. f. wiss. Zool.

Bd. 43. Annales des sciences natnrelles. Serie 3. T. 6. 1846. Handbuch der Gewebelehre des Menschen. 1902. Kolljian : Die Entwicklung der Lymphkndtcken in dem Blinddarm und in dem processus veriniforrnis. Die Entwicklnng der Tonsillen und die Entwicklung der Milz. Arch. f. Anat. u. Pkys. Anat. Abt. 1900. Labeda: Systeme lymphatique : Cours du chyle et de la lymphe. 186(3. Laxger: Ueber das Lymphgefasssystem des Froscbes. Sitzl. d. k. Akad. d.

Wissenseh. Bd/55, 1867, and Bd. 58, 1868. Lauth: Essai sur vaisseaux lyinphatiques. These de Strasbourg. 1824. Lewis, F. T. : The Development of the Vena Cava Inferior. Amer. Jour, of Anat.

Vol. I. 1901-1902. The Development of the Lymphatic System in Rabbits. Amer. Jour, of Anat.

Vol.5. 1906. Cervical Veins and Lymphatics in Human Embryos. Amer. Jour, of Anat.

Vol. 9. 1909. MacCallum : Die Beziehung der Lymphgef asse zum Bindegewebe. Arch, f .

Anat. u. Phys. Anat. Abt. 1902. McCluee : The Development of the Thoracic Duct and the Right Lymphatic Ducts in the Domestic Cat. Anat. Anz. Bd. 32. 1908. The Extra-intimal Theory of the Development of the Mesenteric Lymphatics in the Domestic Cat. Verhandl. d. Anat. Gesellsch. Erg. Heft. z. Anat.

Anz. Bd. 37. 1910. McClure and Silvester: A Comparative Study of the Lymphatico-venous Communications in Adult Animals. Anat. Record. Vol. 3. 1909. Melzer and Adler: Experimental Contribution to the Study of the Paths by which Fluids are carried from the Peritoneal Cavity into the Circulation.

Jour. Exp. Med. Vol. I. 1896. Melzer, S. J. : The Distribution of Solutions in the Cardiectomized Frogs. Jour.

Exp. Med. Vol. 13. 1911. Meyer, A. W. : The Haemolymph Glands in Sheep. Anat. Record. Vol. 2. 1908. Mierzejewskl, L. : Beitrag zur Entwicklung des Lymphgefasssystem der Vogel.

Bull. d. l'Acad. d. Sciences d. Cracovie. 1909. Nuck, A.: Adenographia curiosa. Leidae. 1691. Orth : TJntersuchungen iiber Lymphdriisenentwieklung. Dissertation. Bonn.

1870. According to Chievitz, Orth confused the sympathetic ganglia with the lymphatic glands. Peksa: Studio sulla morfologia e sulla topografia della cisterna chili e del ductus thoracicus. Lab. di Anat. norm, della Univ. di Roma. T. 14.

1908-1909. Polinski, W. : Untersuchungen iiber die Entwicklung der subcutanan Lymphge fasse der Sauger, in Sonderheit des Rindes. Bull. d. FAcad. d. Sciences d.

Cracovie. 1910. Ranvier: Des chyliferes du rat et de 1' absorption intestinale. Sur la circulation de la lymph les petits troncs lymphatiques. A series of articles in the Comptes Rendus de l'acad. des Sciences. 1895-1896. Morphologie et developpement des vaisseaux lymphatiques chez les mammi feres. Arch. dAnatomie microscopique. I. 1897. v. Recklinghausen: Die Lymphgef asse und ihre Beziehung zum Bindegewebe.

Berlin 1862. Retterer, Structure, developpement et fonctions des ganglions lymphatiques.

Journ. de LAnat, et de la Phys. 1901. Sabin : On the Origin of the Lymphatic System from the Veins and the Develop.

ment of the Lymph-hearts and Thoracic Duct in the Pig. Amer. Jour, of Anat. Vol. I. 1901-1902.

DEVELOPMENT OF THE SPLEEN. 745 On the Development of the Superficial Lymphatics in the Skin of the Pig.

Ebenda. Vol. 3. 1904. The Development of the Lymphatic Nodes in the Pig and their Relation to the Lymph-hearts. Ebenda. Vol. 4. 1905. Further Evidence on the Origin of the Lymphatic Endothelium from the Endothelium of the Blood-vascular System. Anat. Record. Vol. 2. 190S. On the Development of the Lymphatic System in Human Embryos, with a Consideration of the Morphology of the System as a Whole. Amer. Jour, of Anat. Vol. 9. 1909. Sala: Sullo sviluppo dei cuori limfatici e dei dotti toracici nelP embrione di polio Ricerche Lab. di Anat. Norm. d. r. Univ. di Roma. Vol. 7. 1899-1900. Sertoli: Ueber die Entwieklung der Lymphdriisen. Sitz.-Ber. d. Wiener Akad.

d. Wiss. Math. Natur. Klasse. Bd. 54. 1S66. Stohr: Ueber die Entwieklung der Lymphknotchen des Darmes. Arch. f. mikr. Anat. 1S89. Die Entwieklung des adenoiden Gewebes der Zungenbalge und der Mandelu der Menschen. Festschr. des Jubil. v. Nageli und Kolliker. 1891. Ueber die Entwieklung der Darrulymphknotehen. Arch. f. mikr. Anat. Bd. 51. 1898. Saxer: Ueber die Entwieklung und den Ban der normalen Lymphdriisen und die Entstehung der rothen und weissen Blutkorperehen. Anat. Hefte. Bd. 6. 1896. Teichmann : Das Saugadersystem vom anatomischen Standpunkte. Leipzig, 1861. Virchow: Die Cellularpathologie. Berlin, 1858.


The fundamental problem in connection with the spleen is its general morphology, — that is, to which germ layer does it belong, — and to this question there is a clear, satisfactory answer in human embryology. The spleen is entirely mesodermal in origin. This was first suggested by Muller (1871) from a study of human embryos. He described the splenic anlage as a thickening of the peritoneum which took place early in the life of the embryo. The spleen arises from a thickening of the dorsal mesogastrium, 1 and is readily made out in embryos of the fifth week, 8 to 10 mm. long. Its relations to the stomach and omentum can be seen in Fig. 515 (after Kollmann, from an embryo 10.5 mm. long) and in Fig. 516 (after Tonkoff, from an embryo 20 mm. long).

In 1889 Toldt advanced the idea that an important part of the mesodermal anlage of the spleen came from the deeper layers of the ccelom epithelium. This has been confirmed and illustrated by Kollmann (1900) and by Tonkoff (1900). All three of them find that in human embryos about 10 mm. long the ccelomic epithelium over the splenic anlage is several layers thick, and that from the 1 Kolliker (1854). Kollmann (1900), Phisalix (1888), Piper (1902), Toldt (1889), Tonkoff (1900).


MUlYliilN Hii\1I5J\1 ULUU1.

deepest layers cells are being transformed into mesenchyme cells. Later the epithelium becomes again a single layer and then this transformation ceases.

The next problem in connection with the spleen is the development of its vascular system. Here it is impossible to give a satisfactory account, for our knowledge is but fragmentary. It will, however, be possible to indicate certain lines of investigation which promise to be fruitful. The study of the vascular system involves

  • "*»


Liver --^O ©i'@:.SSte

Ccelom epithel

Lig. hepato-gastric

Caelom epithel.

Mesog. post."'^= ..-"f^jg'i


affiSk - Mesoderm

4f^g^-%- - Stomach





Caelom epithel.

Fig. 515.

Fig. 516.

Fig. 515. — Anlage of the spleen in the posterior mesogastrium of a human embryo 10.5 mm. long. X 30. (After Kollmann.) Fig. 516. — Diagram of the spleen showing its relations to the stomach and the omentum in a human embryo 20 mm. long. (After Tonkoff.) the use of the injection method, and hence we shall turn to a study of injected pig embryos and fetuses.

In making injections of pig embryos through the umbilical artery, it is striking in how few cases any of the injection mass enters the splenic artery. For example, out of 22 specimens of apparently complete injections made into the umbilical artery, the spleen was injected in only four. Since the injections were all made on living embryos, this is probably due to the relative thickness and power of contraction of the muscle in the splenic artery. To get good injections it is best to open the embryo, tie off the aorta above and below the coeliac axis, and then inject the aorta with a hypodermic needle. When this small length of aorta is well filled, the fluid will run into the splenic artery. In a fetal pig 3 em. long, the entire splenic circulation consists of a capillary network which extends throughout the organ. This condition is maintained until the fetus is 7.5 cm. long, as shown in Fig. 517. This represents the tip of the spleen and shows the central artery and vein which run along the hilum of the organ. As is shown in the figure, the branches of both artery and vein are soon lost in a diffuse capillary network. The branches of the artery can be distinguished for a short distance by being narrower than the veins. Thus the spleen confirms the principle that the primitive circulation of any organ is in the form of a capillary network out of which the arteries and veins are formed. The spleen is characterized by a long persistence of the primitive capillary network.

DEVELOPMENT OF THE SPLEEN. 747 The embryo pig 10 em. long marks the transition stage between this primitive condition and the tj^pe of circulation peculiar to the adult spleen. This point is shown in Fig. 518, where it will be seen that the branches of the central artery and vein extend much farther toward the border of the spleen, and the arterial branches lead into tufts of capillaries, making the anlage of the vascular unit. These capillaries have a wider calibre than those of the preceding stage.

When the fetus is 12 cm. long, as shown in Fig. 519, the transformation of the vascular system has been sufficiently marked to give the key to the adult circulation. By comparison with Fig. 518, which is at the same magnification, it will be seen that between the stages of 10 and 12 cm. there is a rapid increase in size, the spleen more than doubling in width. The position of the central artery and vein allows the comparison. As seen in Fig. 519, the central artery of the hilum gives off a series of branches of the first order which anastomose. These arteries bifurcate into branches of the second and third orders. The branches of the fourth or fifth orders lead into spherules of arterial capillaries which can be seen throughout the spleen, but best at the edge. Most of these spherules have only one artery, but a few receive two ; most of them are isolated, but a few are connected by anastomosing loops. In the upper left-hand corner of the figure can be seen the relation of these spherules of arterial capillaries to the veins. The spherules lead by wide openings into a wide-meshed plexus of venous capillaries, which drain into the still wider venae comites of the arteries of the third order. Thus is illustrated the separation of the artery and vein in the zone of the capillary bed.

At the edge of the organ, it can be seen in total mounts that each spherule of capillaries lies in the centre of a small compartment bounded by trabecular from the capsule. The spherules are the splenic capillaries, characterized by being wider than the usual capillaries. They are at the same time splenic pulp and represent the structural units of Mall (1900). It is the development of these spherules of capillaries that accounts for the rapid increase in the size of the spleen at this time. That they are not accidental is shown by several points,— 1, their constant occurrence at the centre of the lobule at the end of the artery in injections ; 2, their approximately uniform size ; and 3, their definite connection with the veins.

It appears that the number of the structural units of Mall is fixed fairly early; for example, in three fetuses 17 cm. long the number of units along the edge was 150, 204, and 230 respectively, while in three adult spleens the numbers were approximately 190, 180, 260. The size and complexity of these units, however, change greatly; for example, at 17 cm. the average width of a unit is about 0.1 mm. while in the adult it is about 1 mm. Moreover, the embryonic unit consists of one central artery with a single bunch or spherule of capillaries leading to the vein, while the adult unit, as shown in Mali's Fig. 1 (Amer. Jour, of Anat., vol. 2, p. 321, 1902-1903), consists of a central artery with branches which end not in a single spherule of capillaries but in clusters of capillaries like a bunch of grapes. These clusters are the splenic capillaries or pulp. One of the spherules on the edge of the spleen in Fig. 519 shows the bifurcation of the central artery; how this complex unit of the adult is made out of the simple one of the embryo is yet to be determined, but the evidence of embryology is that the capillaries of the spleen are of a wider calibre than the usual capillaries, and that the wider capillary bed is compensated for in the development of the musculature of the spleen by which the capillary bed can be emptied with ease.

Mall (1902 to 1903) proved, by the method of subjecting the living adult spleen to a variety of injection methods, of which the most crucial test was the fixation of the spleen by the injection of formalin into the living animal, that



the splenic pulp is the capillary bed of the spleen; that the pulp intervenes between the artery and vein, and, in the normal, living animal, is engorged with blood when the spleen is hypersernic. Hence the capillary circulation of the adult spleen is a cavernous one. The point at which the type of circulation of the early fetal stages — namely that of a primitive system of closed capillaries like the rest of the vascular system — changes over into the secondary type of circulation of the adult, namely a cavernous circulation, is shown in Fig. 519, for the fetal pig.

The work of Mollier (1911) carries us a step farther, by his beautiful specimens of the structure of the wall of the splenic veins. He also, as Mall had done, overthrows the idea of a homogeneous membrane around the endothelium of the adult splenic veins. He shows that the wall of the splenic veins consists of a reticular syncytium of endothelium with denser masses of protoplasm around the nuclei and wide-open meshes between.

It has thus become clear that there is a direct pathway from the terminal artery through the cavernous capillary system into the venous sinuses, and that





Fig. 517.

Fig. 518.

Fig. 517. — Piece of a total mount of an injected spleen of an embryo pig 7.5 cm. long, showing the capillary plexus which is characteristic of the circulation at this stage. X 47. A., artery; C.a., central artery of hilum; C.v., central vein of hiluni.

Fig. 518. — Piece of a total mount of an injected spleen of an embryo pig 10 cm. long, showing that the anlage of the splenic unit is a tuft of widened capillaries. X 47. A.b.c, anlage of capillary spherules or units; C.a., central artery of hilum; C.a.l., central artery of lobule or unit; C.v., central vein of hilum.

to investigate the nature of the endothelium of the cavernous capillaries or splenic pulp is a problem which is becoming more and more hopeful. The transition stage between the two types of capillaries is the next point of attack.

The third problem in relation to the spleen is in connection with its function of the formation of red blood-cells. Kolliker (1854) was the first to suggest the idea of the spleen serving as a place for the formation of erythrocytes and leucocytes, and called attention to the relation of the giant cells to red-blood formation. Luzet (1901) pointed out in human embryos of from 3 to 5 months there were more nucleated red blood-cells than in the heart's blood. Sophie Lifschitz (1906) has shown, in a work which I regard as important, that the active formation of red blood-cells takes place in fetuses between 15 and 30 cm. long. She plotted the curve both of the number of red blood-cells and of the giant cells of the spleen, and found that both increase together to a maximum in fetuses 18 cm. long, while the curve



almost at the zero point at 30 cm. the nucleated red blood-cells form cells. It may be noted that this formation corresponds with the period of the formation of the capillary spherules or spleen pulp, which makes the spleen fall in line with the recent work on the bone-marrow by Bunting (1906), and on bone-marrow in the kidney of Maxhnow (1907), that red blood-cell formation

decreases at 2-1 cm. and is Moreover, she noted that clusters around the giant period of red blood-cell



Fig. 519. — Piece of a total mount of an injected spleen of an embryo pig 12 cm. long, showing the vascular units. X 47. The arteries are shown darker than the veins. A.b., anastomosis between two capillary balls; c.a., central artery of hilum; c.b., arterial capillary ball; c.b., central artery of hilum; T 1, vein of first order; this vein up to the point marked X was taken from another area of the same slide where the injection was more complete; it is the rule that the veins accompany the arteries in this manner; V.C3., vena comites of the third order; v.c.p., capillary plexus of veins.

goes on within the capillary bed, that it is intravascular. Lifschitz (1906) also called attention to the rapid increase in size of the spleen during the period of red blood-cell formation, which Was also noted in connection with the period of spleen-pulp formation. The injection experiments show that its meaning lies in the rapid increase in the size of the capillaries. The fundamental point that the early spleen is only undifferentiated mesoderm, and that this condition remains until the embryo is 7 cm. long, was noted bv Van der Stricht in 1892. He states that there is a


primitive stage in which the structure of the spleen is more or less uniform before there is any differentiation into splenic pulp and Malpighian corpuscles; then he finds a transitional stage characterized by an increase in white corpuscles in certain areas, an increase in erythroblasts and a retardation of the circulation, and finally a secondary stage in which the adult organization of the spleen with pulp and Malpighian corpuscles is established. The fourth problem in the development of the spleen is in connection with the ellipsoids and Malpighian corpuscles. It is definitely known that these structures belong to the latter half of fetal life. In the spleen, as has been shown in the lymph-glands, the lymphocyte first appears in the adventitia of the artery, so that, though red blood-cell formation is within the capillary bed, the lymphocyte is extravascular. The ellipsoids, or capillarhulsen of Schweigger-Seidel, which are on the course of the smallest arteries, appear, as Bannwarth (1891) has shown, before the Malpighian corpuscles, which "are the round follicles along the larger arteries. He found the ellipsoids in a four-months human fetus, whije later — that is, in a seven-months fetus — the ellipsoids had disappeared and follicles were present. The follicle is found only in mammals and some birds, while the ellipsoids occur in fishes as well, as has been shown by Whiting (1895). Thus the ellipsoid is the primitive lymphoid structure of the spleen. In describing their development, Bannwarth shows that in the spleen, as in the lymph-gland, leucocytes appear in the loose adventitia of the artery, and at the same time there is a development of this adventitia, by which the connective-tissue fibrils are laid down in more or less concentric rings around the artery, forming the delicate reticulum characteristic of the lymph-follicle in generaj. Thus, the spleen, hemal glands, and lymph-glands are all vascular structures and are all built on the following simple plan : 1, along the arteries are clumps of lymphocytes in a reticulum called ellipsoids or Malpighian corpuscles or follicles; the ellipsoid is a special name used in the spleen for the oval masses of lymphocytes which lie nearest the capillary bed ; 2, the capillaries, whether they be lymphatic capillaries in lymph-glands or blood-capillaries in the hemal nodes or the spleen, are all wider in calibre than other capillaries. They are densely packed together, and have been termed either sinuses in lymph-glands or pulp spaces in the spleen.


Bannwarth : Die Milz der Katze. Arch, f . mikr. Anat. Bd. 38. 1891. Bezaxpon et Labbe: Trait e d'hematologie. Paris, 1904.

Bunting, C. H. : Experimental Anaemias in the Rabbit. Jour, of Exp. Med. Vol. 8. 1906.

DEVELOPMENT OF THE SPLEEN. 751 Choronschitzsky, B. J. : Die Entstehung der Milz, Leber. Gallenblase, Baueh speicheldi*iise und des Pfortadersystems bei den verscbiedenen Abteilungen der Wirbeltiere. Anat. Hefte. Bd. 13. 1900. Daiber, M. : Zur Frage nacb der Entstebung und Regenerationsf abigkeit der Milz. Jenaisebe Zeitscb. f. Naturwissensehaft. Bd. 42. 1907. Janosik, A.: Bemerkungen zu der Arbeit, Dr. W. Tonkoff; Die Entwicklung der Milz bei den Amnioten. Arch. f. mikr. Anat. Bd. 57. 1901. Kolliker, A. : Mikroskopisebe Anatomie. Zweite Halfte. Leipzig, 1S51. Kollmann, G. : Lebrbucb der Entwicklungsgesckiehte des Menscben. Jena, 1898. Die Entwicklung der Lyinphknotchen in dem Blinddann und in dem processus verrniformis. Die Entwicklung der Tonsillen und die Entwicklung der Milz. Arch. f. Anat. u. Phys. Anat. Abt. 1900. Lifschitz, S. : Leber die Entwicklung der embryonalen Milz. Med. Diss. Zurich, 1906. Luzet: Etudes sur les anemies de la premiere enfance etc. These. Paris, 1891. Mall, F. P. : The Lobule of the Spleen. Johns Hopkins Hospital Bulletin. 1898. The Architecture and Blood-vessels of the Dog's Spleen. Zeitschr. f. Morphol.

u. Anthropol. Bd. 2. 1900. The Circulation through the Dog's Spleen. Amer. Jour, of Anat. Vol. 2.

1902-1903. Maximow, A. : Experimented Untersuchungen zur postf etalen Histogenese des myeloiden Gewebes. Beitr. z. path. Anat. u. allgem. Path. Bd. 41. 1907. Mollier. TV: Leber den Bau der kapillaren Milzvenen (Milzsinus). Arch. f.

mikr. Anat. Bd. 76. 1911. MiiLLER, YV : Milz, in Strieker's Handbuch. Leipzig, 1871. Neumann, E. : Neue Beitrage zur Kenntnis der Blutbildung. Arch. d. Heilkunde.

Bd. 15. 1S74. Ueber Blutregeneration und Blutbildung. Zeitschrift fiir klinische Medizin.

Bd. 3. 1881. Phisalix, C: Etude d'un embryon humain de 10 mm. Arch, de zoolog. experim.

et gener. T. 2, p. 6. 1SS8. Piper, H. Die Entwicklung von Leber, Pancreas und Milz. Inaug.-Diss. Freiburg i. B. 1902. Sabik, F. R. On the Development of Lymphatic Nodes in the Pig. Amer. Jour.

of Anat. Vol. 4. 1905. Sassuchin, P. : Leber die kindliche Milz. Inaug.-Diss. St. Petersburg, 1899. Schmidt, M. B. : Ueber Blutzellenbildung in der Leber und Milz unter normalen und pathologischen Verhaltnissen. Ziegler's Beitrage. Bd. 11. 1892. Toldt, C. : Zur Anatomie der Milz. Wiener klin. Wochenschrift. 1S89.

Die Darmgekrose und Netze im gesetzmassig und gesetzwidrigen Zustand.

Denkschriften d. k. Akad. d. Wissenschaften zu Wien. Bd. 56. 1SS9. Tonkoff, W. : Die Entwickluns: der Milz bei den Amnioten. Arch, f . mikr. Anat.

Bd. 56. 1900. Van der Stricht, Omer : Xouvelles recherches sur la genese des globules rouges et des globules blanc du sang. Arch, de Biologic T. 12, p. 199. 1892. TVoit. O.: Zur Entwicklung der Milz. Anat. Hefte." Bd. 9. 189S. TVhitixg. A. J.: Comparative Histology and Physiology of the Spleen. Trans.

Royal Society of Edinburgh. Vol. 38. Part. 2. No. 8. 1895.